stagecraft examples: the release gate says which SHA goes in commit and which in mergeCommit
stagecraft: start on a claimed key takes the claim's externalRefs, and a terminal work item can be given its ticket
SWAMP_SERVER_TOKEN does not authenticate with --server, while --token and SWAMP_SERVER_TOKEN_FILE do
Add-on methods (export const extension) are not found when a step is placed on a worker
stagecraft: a reconcile view that cross-checks reality (git ancestry / registry) against work-item and ticket status
A hung stagecraft publish holds a tracker's datastore lock 15+ min; after the holder dies the lock stays orphaned and run doctor reports the datastore healthy with no way to release it
serve: the cron-fire dedup claim has no deadline, so a hung control plane stalls every scheduled fire
serve: a capability probe answered by a follower does not speak for the leader that handles the forwarded request
data prune through serve deletes data of an auto-definition whose declared type is invalid
A serve host with no root in host.yaml makes its root an ordinary workspace, registered like any other, with serve's default auth
extension rm misses dependents that pin the removed extension as name@version
workspaceHub app-socket integration test expects the child's echo before the hub's own refusal
Team mode: a follower serve's commits.db WAL is never checkpointed and grows about 1 GB a day (1.7 GB after 2 days), and filled the disk
Team mode: a booting serve blocks its event loop for 21-51 s on a large datastore, longer than the writer lease's 25 s lapse, so a standby can take over mid-boot
swamp workspace import silently takes a serve host's workspace offline when its folder or name is in the host's workspaces folder
Remove the serve host's host folder layout and --host-dir: the host always reads host.yaml from the config dir and serves the registry's workspaces
serve host: a host's own workspace is suspended on an unreadable wanted file on macOS (folder spelling), and a workspace hub test assumes reply order
@swamp/cfgmgmt firewall/apt_repository/ssh: follow-ups from #4840 verification review
stagecraft studio: the default look passes axe and WCAG AA for graphics (journal tick contrast, graph tree ARIA, panel keyboard scroll)
Hosted workspace create answers host_error, not already_exists, when a plain file holds the name in the workspace location
A host-created workspace's wanted file records the repo it was made for, so a new repo at the same path is never the host's own
serve HTTP action follower test flakes: one person's rate limit through a follower can fail with 'the caller reached the limit'
The first swamp serve host run creates host.yaml in the config dir with reasonable defaults instead of exiting
Forms submit natively before their island hydrates, reloading the page and wiping what was typed
A serve host with no host folder creates workspaces in the workspace location
ensureReadSet downloads every text resource of a referenced model, not just the instances an expression reads
serve resolves realpath for thousands of files under .swamp/outputs/swamp/worker/set_status
stagecraft studio: a repository can define a theme (colours, fonts, corner rounding) for its studio
swamp serve host runs without a host folder: host.yaml in the config dir and its state in the data dir
stagecraft: a read of what the current stage expects (prompt, products and their fields, gates)
Streaming resource writes from workers through the serve data plane
Streaming resource writes from behaviors with no single-string size ceiling
data query --select reads only the content paths the expression needs
Workflow steps parse every written resource in full, and past the V8 string limit their outputs are silently empty
Flaky test: serve host registry 'a repo imported while the host runs is listed stopped...' reads no status file after the workspace already answers stopped
Flaky test: startLoadedWorker 'a Worker that loads at once starts once and answers' times out loading its fixture under load
@swamp/cfgmgmt firewall: placing a deny with first or before can lock out the run's own SSH session
@swamp/cfgmgmt: SSH ControlMaster sockets and pins ignore TMPDIR (hard-coded /tmp)
@swamp/cfgmgmt apt_repository: gpgKeyUrl writes keyring with umask-dependent perms (640 on umask-027) -> apt signature verification fails
cfgmgmt firewall (iptables): rule placement with before, after and first
stagecraft: factory definitions can take prompts and skills from repo files
A v3 write command waits the full sidecar start deadline after the supervisor it launched has already exited
Oversized resources are accepted silently at write time, then crash later runs with a V8 out-of-memory and no diagnosis
DataQueryService.refreshTypes test fails: its interleaving reader never runs during the refresh
stagecraft: a dispatch's product write records its usage, usage stays complete across subagents, and onBehalfOf is recorded consistently
Registered workspaces: a workspace registered anywhere on the machine is listed on the serve host and served in place on resume
Docs: resourceSize limits in .swamp.yaml and data query --sort size
CEL model.<name>.resource loads every latest resource of the behavior, not just the one referenced
data query --select content and foreign-content hydration load oversized items without the resource load guard
Test flake: v3 replication 'an owned sidecar whose owner's pid names another process stops' fails with 'the Litestream sidecar did not start in time' under load
stagecraft studio: Inspect shows each stage as a field explorer, not YAML
Add built-in JSON query support to the Swamp CLI
Extension push: a registry 5xx reads as the user's fault and gives no next step
`swamp data query --json` fails with "Invalid string length" once results exceed V8's maximum string length
stagecraft: dispatch prompts read injected products by latest record, not the version the run indexed
stagecraft DESIGN.md: say the stored-as-received check catches only a redacted secret, and --literal-inputs is what stops evaluation
stagecraft skill: gate summaries at the decision point, systemPrompt on workflow and behavior stages, claim's key, in-flight definition changes
Flaky test: workspaceHub a busy child is not stopped until its runs end times out waiting for the hub to ask the busy child again
extension push: content hash still depends on the extension's folder with paths.base: manifest (review report reported missing)
Flaky test: DataQueryService.refreshTypes re-reads only the given types fails intermittently when its 0 ms timer never fires during the refresh
swamp serve data.query builds its whole response as one WebSocket frame, so a remote query past V8's string limit fails on the server
swamp data get --json fails with Invalid string length on one record whose pretty-printed content passes V8's string limit
v3 bucket writer intermittently refuses with a false 'another user or host runs this datastore's Litestream sidecar' when a stale presence record from another kernel or pid namespace meets its own just-launched supervisor
Store large values once: content-addressed sharing across resources and versions
Bounded-memory access to large resources from behaviors: partial reads and streaming writes
Work-item commands run from the wrong repo report the work item as not started
stagecraft: an empty, never-started instance takes a work item's key, so claim falls back to <key>-<tracker>
swamp data get and data versions report Behavior not found for data whose behavior definition was deleted
Scheduled flaky-test detection is failing
The serve host takes its workspaces from the workspace registry, with no change to what it serves
Tailnet: developer machines can reach each other's tailscale serve by default
A serve host registers the workspaces it runs in the workspace registry, where they are
stagecraft: publish to the bound tracker after each write, and infer workItem
Team mode: a far member's large-output save takes about 2.5x v2, all in the single-connection upload to serve
Docs: the serve workspaces design keeps one list of workspaces, the registry, and a host needs no host folder
Test flake: v3 replication 'an owned sidecar whose owner's pid names another process stops' fails with no control socket under load
Flaky test: serve host 'a delete interrupted before or during the removal finishes at the next start' reports a never-announced workspace as removed
Scheduled flaky-test detection is failing
stagecraft: writes print a short status by default; full output on request
Workflow pre-validation refuses a method argument supplied as a global argument when the definition authors a methods.<method>.arguments block
Flaky test: hosted workspace lock test loses its own lock under load
The CLI signs in to a hosted workspace through the root, after one login to the host
workflow validate and run re-walk models/ once per distinct behavior name in nested workflows (follow-up to #4903)
The dashboard switcher lists a serve host's workspaces, read from workspace.list on the root
Dashboard sign-in under a workspace segment: under a serve host the page signs in to the workspace's serve
A dashboard session cookie per path: a serve with a path in its address names and scopes its cookie for that path
workflow validate and workflow run re-scan models/ once per distinct step behavior (O(steps x definition files) file opens)
behavior validate re-scans models/ once per distinct expression model reference (O(references x definition files) file opens)
@swamp/ssh: repin fails on Windows targets - POSIX `true` liveness probe not found in cmd.exe, new pin never persisted
Docs: team serve rolling restart runbook, leader takeover time, and how clients treat an inFlight refusal (#4811, #4824, #4830, #4839, #4853)
Docs: how to rotate swamp serve's TLS certificate (rolling restart on Linux, or TLS at a proxy)
Docs: a planned swamp serve restart hands in-flight runs over instead of interrupting them
Docs: swamp serve restart behaviour from #4566 (/health draining body, --reuse-port, --shutdown-listener-grace, startup not_ready refusals)
Admins aren't pinged when a profile showcase is submitted for approval
Admin panel can grant a Personal License to a collective
serve --reuse-port on macOS: the AddrInUse refusal blames 'a process listening there without --reuse-port' when the cause is the OS
workflow run --server: after reattaching to a run continued across a serve restart, the step in progress shows the reattach time as its start, so its duration is understated
swamp behavior list re-scans the extension model-definition tree once per distinct type (O(types x extension-definition-files) file opens)
Team mode: a serve booting with a cold page cache takes 109-122 s to become ready on a real-size datastore (reads about 1 GB)
A host's router turns a child's refused WebSocket into 502, hiding why the workspace or root refused it
Ship a linux-riscv64 binary (enables real swamp in the browser via unkernel)
serve refuses shell steps that splice vault.get into run on protocol 5 workers (shell_data_placement_refused)
serve hangs forever if its shutdown throws before the listener closes
serve --json --exit-on-stdin-close never exits when its host is SIGKILLed while a run is in flight
Scorecard scanCode: scan each code module once with the extension-scoring FromSites functions
serve --json never counts webhook (and likely scheduled) run outcomes in /api/v1/health metrics
workspace.list on a host root refuses non-admin callers signed in through the CLI
serve SIGTERM skips the failed-gated cleanup job of a run the user cancelled before the shutdown
serve --json writes {"status":"stopping"} to stderr instead of stdout while a run is in flight
extension-scoring: markdownImports parses every code file with Babel uncapped, and writtenSpecifiers parses them again
The sign-in exchange: a root token becomes a workspace token through a single-use code
Investigate replacing @babel/parser with oxc for parsing extension source
A workspace kind in grants: a group can be given its own hosted workspaces on the root
CLI: idempotent swamp-club requests don't retry a 503, 502, 504 or timeout, so one blip fails the push registry checks
update.restart-daemons restarts a pre-existing worker daemon with no stable cache dir into an enrollment loop
Serve links: accept and describe workspace selectors in a consented ceiling
swamp serve workspace commands: create, list, get, suspend, resume and rm
A worker daemon enabled without --cache-dir re-enrolls as a new machine on every restart, so a single-enrollment token stops it
Workspace requests on the root: serve handles workspace.*
Enrollment-token methods run from a workflow step in serve bypass the gateway's token write ordering
Default off: the serve host starts a workspace's serve only when its wanted file says active
Extension scoring blocks the API event loop for over a minute on large extensions, causing 503s on the scoring pod
Docs: enrollment token rotation now reaches workers from workflows, schedules and local rotate
Workspace services: create, list, get, suspend, resume and delete a hosted workspace
Step-lease and pending-dispatch methods run from a workflow step in serve bypass the dispatch tail's per-record ordering
serve: a follower that does not name its leader tells an inFlight serve_not_leader caller to try again
java4shell: add Windows scanning for parity with the other CVE extensions
Docs: autoupdate can restart worker and serve daemons, and set up non-interactively
java4shell 2026.11.18.1 upgrade breaks behaviors created on 2026.11.17.1
serve: a leader's serve_not_leader raised after a run start executed carries no inFlight, so clients may start it again
java4shell: recognize vendor fixed builds shipped after disclosure
workflow placement: run a workflow transparently on a specified host, local or via swamp workers (incl. nested child workflows)
workflows: a nested workflow step should inherit the caller's placement when the child declares none
sub workflow placement on workers breaks some of my extensions my
Source-run update can replace the Deno runtime
Swampify: a Preview badge on the swamp start card, and the opt-in env var goes
Add @swamp/cve/java4shell scanner for CVE-2026-114938 (Java4Shell)
Docs: nested workflows inherit their caller's placement (swamp-club#4766)
client: an HTTP action refused as serve_not_leader with inFlight is retried and can run twice
Stopped: Team mode: a member's v3 copy keeps growing under steady load even with follow-serve's-compaction (about 16 MB/h, no plateau)
Team mode: faster takeover, piece 1: a shorter writer-lease TTL (then the follower's physical copy)
Wanted and status files: suspend, resume and delete a hosted workspace
The serve host passes an OAuth client from host.yaml to its children
workflow approve: add --resume to continue the run once all gates are decided
Flaky test: webhook held delivery to a starting leader counts a third forward under load
Docs: document swamp workflow approve --resume in the manual
@swamp/cve/http2shell: old mainline RHEL 9 libbrotli z-stream builds rated check-vendor instead of vulnerable
swamp serve host: one listener that runs a serve per workspace folder
Test flake: team forwarding 'a leader that left two forwards in a row unanswered is refused at once' sometimes forwards and fails instead of refusing
Flaky under load: team forwarding test 'a leader that left two forwards in a row unanswered is refused at once' gets a may-have-run reply
Flaky under load: workflow report race, team-mode read-lock and recovery chunk-read integration tests
Tests no longer depend on Deno's npm resolver, whose TaskQueue can lose a wakeup
test: the scheduled flaky-test job skips the Litestream sidecar tests
Process-wide module state in the v3 datastore forces tests to isolate actors in Workers
test: three 'crash' tests check SQLite rollback, not crash recovery
test: test Workers that only isolate a writer lease or one command can run in the test's isolate
Scheduled flaky-test detection is failing
cve/dirtyfrag: data names use the raw host or container label, with no sanitising or clash check
A follower that forwards a v3 datastore chunk upload to its leader may get 411, because the forward drops Content-Length
swamp serve host: give each hosted workspace its own environment variables
serve: a team serve restarts one host at a time with no failed requests and no lost runs
cfgmgmt firewall (ufw): an input for rule position (insert before/after)
cve extensions: data names can still collide or be refused on Windows (trailing dots, reserved device names, case)
workflow evaluate: guard on a forEach step is over-resolved to a boolean and fails schema validation
Load serve's subcommands lazily through a serve route table
Manage the swamp-club profile from an extension behavior: a profile API and a token scope
Swampify page doesn't show new candidates or the offer while the tab stays visible
The serve host supervisor: start, watch and stop one serve per workspace
data query --select crashes in log mode when a list projection fails for one record
Structured data query content renders as an unreadable table
The workspace router: forward /w/<name>/ to that workspace's serve
S3 and GCS datastores: copy replica files into a new Litestream lineage (the replicaPaths copy method)
serve shutdown: a nested child run cut off by shutdown ends cancelled without its cleanup, and the parent's resume starts a new child
Swampify: guide the last mile from built to everyday use (an adopted state)
Team mode: a member's copy follows serve's compaction by installing serve's base per partition (small partitions whole, hot ones by difference), so copies stay bounded without a local rebuild
serve: a run left RUNNING after a restart is never seen as stale when the new serve reuses the old PID (containers)
Team mode slice 2: cold-read catch-up writing, catch-up before the lock, reclaiming a dead predecessor's lease, a failed lock-time catch-up goes offline, batched offline flush, and serve boot's running-runs scan
serve: swamp serve reload rotates the TLS certificate, key and CA certificate without a restart
Lab issues page: toolbar overflows and clips controls when list pane narrows
ManagedServe: the entity that decides when a hosted serve starts, restarts and gives up
serve: a planned restart parks runs at a step boundary and the new process continues them
@swamp/cfgmgmt: there is no way to require pinned host keys; a model without `nodeHostKeys` trusts the first key it sees
serve: during the shutdown drain, an aborted scheduled run keeps executing its remaining steps
Every serve client works behind a path prefix
serve: its own access reload is read as a foreign change, the whole data catalog is invalidated, and the rebuild blocks the event loop
The worker's data plane keeps the connect URL's path
test: the command routes test still stalls at its 120 s worker deadline when it runs in its own deno test process
stagecraft studio: a light theme that follows the system, with a toggle
serve --exit-on-stdin-close: a serve drains and exits when its stdin closes
Dashboard data Behavior filter includes workflow artifacts
Dashboard audit prompt references missing default config file
v2 S3/GCS: a mutating run that can't confirm its model lock at push exits 1 but keeps its data, and the next sync publishes it to every clone
Test flake: LitestreamReplicator.quiesce 'refuses a sidecar it cannot verify' sometimes refuses for a busy sidecar instead
Test flake: auto-continue 'a run aborted for its maximum duration records the cancel before the abort' sometimes finds no cancel reason
A step that depends on an empty forEach starts before that forEach's own dependencies have finished
Flaky: replication test 'an owned sidecar whose owner's pid names another process stops' reads a presence record the sidecar already removed
Test flake: v3 replication 'an owned sidecar whose owner's pid names another process stops' reads presence.supervisorPid of an undefined presence
Test flake: fan-out rollout 'maxFailures 25% over 8 lanes halts on the third failure mid-batch' sometimes dispatches one lane more than expected
Scheduled flaky-test detection is failing
Flaky: command_routes_rules_test's completion-callback test gets no answer from its full-tree worker within 120 s in verify-build
Team mode item-4 design note: a copy follows serve's compaction, never falls behind it, catches up by difference, recovers through an undo log, and bounds a full rebuild
Flaky: datastore_v3_replica_verify_test times out waiting for the sidecar's verdict when restarting a broken lineage under a live sidecar
Test teardown killSidecars throws EPERM when the Litestream group has already exited (macOS)
ServeLock test 'a renewed fenced lease keeps its fence and floor' times out waiting for two renewals under load
Team mode: a member's or serve's copy is never slow to become usable (recovery, serve boot, catch-up after idle or compaction, login reset) — cost proportional to change, never to datastore size
Team mode failover: one member stalls for minutes in most rounds, takeover runs past 60 s, and a serve outage exits slowly with exit 1 (run 7)
serve auto-continuation follow-ups: stale claim comment, cancelled continuation audited as failure, failed recover leaves its claim
validate: spurious ambiguous-exit warnings for evidence absent in this stage/cycle, and for complementary cel comparisons
Docs: automatic continuation of interrupted runs in swamp serve (autoContinue)
@swamp/cfgmgmt firewall: pre-check edge cases (firewalld undo and old-version detection, iptables inherited comments)
@swamp/ssh: `timeoutSec` is not honoured when ControlMaster is on
@swamp/ssh: a model that calls a fleet can't tell how that fleet checks host keys
@swamp/ssh: a model that calls a fleet can't tell whether two fleets reach a host by the same route
@swamp/cfgmgmt systemd: a failed enable or start does not fail the step
@swamp/cfgmgmt firewall: removing the rule that lets the run's own host in cuts the run off, with no deadline
@swamp/cfgmgmt firewall: no syntax check of a rule before it is applied
Registry: an extension earns "verified repository" and an A scorecard while its source is not in that repository
@swamp/cfgmgmt apt: installing on a fresh machine fails until a separate refresh
@swamp/cfgmgmt user and group names are not validated
@swamp/cfgmgmt systemd: a unit cannot be removed (no ensure: absent)
agent-sessions assess: a Claude Code session reachable at two paths is counted twice
Swampify: compare counts model steps, fixes and the latest run, and the skill records built only for output the user accepts
Swampify: comparison v3 records the latest run, build sessions, unrecorded model usage and occurrences with no calls
swamp start: Codex can run swamp inside its sandbox, and a 421 names --trusted-hosts
Swampify: start offers windows up to all history, you choose where your agent opens, and start and the dashboard share one breadcrumb
Let a revoked worker token's name be reused (or delete revoked token records)
swamp serve: an admin "restart" that ends the serve and lets its in-flight runs be continued
swamp start: finds more Linux terminals (Ptyxis and others), a failed launch says what it tried, and resume reopens the conversation start launched
Swampify: a probe users run and share privately so we can add other agents, and the skill reads windows up to all history
agent-sessions assess: fails on transcripts over ~512 MB and skips symlinked transcripts
workflow resume ignores globalArgs of direct type execution steps; auto-definition created with empty globalArguments
Scheduled flaky-test detection is failing
serve: automatic continuation follow-ups (scheduler rebuild overlap, reconciliation retry, nested refusal re-judge)
serve: continue an interrupted run automatically when every step the restart cut off is safe to run again
Extension versions: print one spelling in log output (drop the v prefix in list, outdated, update, search)
serve: add a fitness test so no serve route is added under the /hooks/ webhook route-file prefix
Docs: extension management reference shows v-prefixed versions in extension list and update output
Docs: show the guard result on guarded skipped lines in make-workflows-idempotent
Docs: reaching swamp start and swamp dashboard from another machine (--host)
Docs: vault manual says legacy vault type names are automatically mapped, but vault create and migrate refuse them
Document trusted telemetry endpoint and authentication semantics
GCP Drive files enrichment: download_content can overwrite the state written by upload
vault: legacy type deprecation warning wraps names in doubled quotes
Support image attachments in issues (Lab and CLI)
GCP codegen: StateData inferred from StateSchema may hit TS2589 now that nested state objects are .partial()
GCP codegen: generate media upload/download from Discovery mediaUpload and supportsMediaDownload
v3 targeted recovery leaves a different data/<model>/<name>/latest than a full rebuild after a rename onto a deleted name with an interrupted writer (deterministic property-test seed)
workflow history get: show the username for who decided an approval, not just user:<id>
Docs: dashboard and history get name who started and decided a run by username
swamp skill references still point to separate skills that were merged into the swamp skill
swamp start: make Swampify opt-in while it is experimental
swamp skill vault references are out of date: JSON output examples, built-in vault types, and a behavior that does not exist
worker token rotate counts a machine as retired after it took an earlier rotation in-band and disconnected
Swampify: after a swamp start kick-off, the skill hands the pick to the Swampify page and keeps one build per session
Swampify: start page card at the top, Swampify page refreshes on return, and briefs link to the page
Flaky: SqliteReadModel.recover tests depend on wall-clock timing and fail under load
Flaky: command routes test times out when a batch of route workers stalls importing main.ts
local_encryption silently generates a new key when .key is missing, making every existing secret unreadable
start_server_test: the one-interface-address test flakes when a parallel test binds 127.0.0.1 on the same port
Docs: rotating and revealing worker enrollment tokens in the worker-fleets guides
local_encryption vault base_dir does not expand $HOME, unlike datastore.path
Supported way to rotate a connected worker's enrollment token over its own connection
swamp start --no-browser prints out URL on localhost
Swampify: the skill recognises both swampify briefs, and compare measures raw against swamp in a compared state
Swampify: swamp start offers "Start from your past sessions" and launches a swampify kick-off task
stagecraft: rebind a work item to another tracker (mirrored built-in tickets to Jira)
stagecraft: a factory-wide read of work items waiting on a person (no studio, no per-item status loop)
Swampify: read-only panel on the dashboard's Agent sessions page over core's contracts, and agent-sessions.md made current
Team mode: recoveries still take tens of seconds on members and 6-16 minutes on a serve after a kill (targeted path slow; the leader's dirty set forces full rebuilds)
ABAC support in @swamp/aws-sm
swamp extension update writes an upgraded extension's skills into a skill directory another extension owns
datastore push: a dropped push after a reported-success run isn't persisted (run still exits 0); a wedged push holds the datastore lock until the process is killed
Docs: swamp serve reload now applies scheduling, the scheduled-run limit and timed-service on/off
Docs: manual page for swamp workflow schedule list and pause
Docs: catalogue the audit actions swamp serve records
Docs: document garbageCollection.autoDefinitions and the run gc changes from swamp-club#4299
Docs: team members' CLI retries reads refused during a serve failover, and the new recovery failure code
Docs: swamp serve reload now applies most serve.yaml settings without --hot-reload or a restart
Docs: hot-reload manual page lists which serve settings reload, including grant settings and admin removal
A dense schedule fires nothing during the repeated hour after a DST fall-back
serve: telemetry for integrations: report hosted integrations to swamp-club, a client kind on usage events, and events for approvals and run control
stagecraft: studio state_test 'after its event stream reconnects' is timing-flaky under load
serve: let methods act on swamp-club as the user who started the run (on-behalf-of token exchange)
serve: shutdown does not wait for or cancel the own-cron replay that follows a scheduler turn-on
serve: a webhook or cron run aborted by shutdown ends cancelled instead of interrupted, so it cannot be recovered
A schedule armed during a DST fall-back hour fires continuously in serve
swamp-club: store the integrations each serve reports, and send serve-link events to analytics
Serve won't release a worker token's secret to a dispatched step (reserved_secret), and this isn't documented
@swamp/video: generate intro and bed music sized to the take (ElevenLabs first; Suno once its API is open)
data.latest(): catch a name that equals its spec name before run time, and say how to fix it
Schedule catch-up counts slots before the watermark across a DST fall-back in the host timezone
Workflow guard does not see data written by an earlier step in the same run (data.latest evaluates empty)
@swamp/video/obs: add the brand stage to an existing OBS scene, with the webcam ring and mask optional
Rename a behavior type (e.g. move it to another namespace) without losing its data
@swamp/video: describe and chapters for a joined video
Studio Board orders stage columns by position within each tracker-status lane, not by process order
serve: a rolling restart's boot sweep expires the leases, dispatches and workers the old process is still using
swamp behavior type rename: move a behavior type's data, outputs and catalog rows to a new type name
A corrupt extension lockfile prints a stack trace instead of a clean error
Docs: datastore-configuration vault.get section says extension vault types must be pre-installed
Team mode: a single save is slow on GCS (1.7 s near, 5.3 s far p95), mostly waiting for a lock nobody holds; far workflows and the first command after idle too
Dashboard: show the client (e.g. integration:slackwater) on activity events and run details
Datastore fails to open on fresh checkout: early vault service never auto-resolves the datastore-auth vault's extension type
swamp issue: choose the target collective and control issue visibility -- at filing time and afterward -- from the CLI (with a settable default)
Render <details>/<summary> in serve's report view
Let a type mark a method argument as data, and read an input file's content as data
An invalid --log-level value fails with a raw TypeError and stack trace instead of a usage error
Swampify: @swamp/agent-sessions adopts core's contracts, and its skill reads past sessions with the user's chosen model, records the offer and pick, hands the pick to getting-started
Root-help property test in command_routes_property_test.ts fails when help is the value of --log-level
Lazy loading L8: load only the invoked subcommand in extension, vault, datastore and access
swamp-club: a link request withdrawn by a plan lapse is stored with the secret-rotation reason
A crashed FileLock holder's stale lock can be reclaimed by two waiters at once, so two runs hold one model's lock
stagecraft artifact digest breaks when artifact content holds a resolvable ${{ }} expression (interpolated on read)
Manual: build your own integration with swamp serve, using Slackwater as the reference
Swampify: contracts move into core and are revised for their readers (candidate set v2, offer v1, comparison v2)
serve: close three timing windows in the scheduler reload lifecycle (boot reload, shutdown during turn-on, stale remote own-run replay)
Lazy loading: root --help renders from route metadata instead of loading every command
swamp-club: serve-link copy, the username step, switching account, and withdrawn-link copy, in one pass
Lazy loading L7: ship the bundled build (bundled by default, tested on macOS and Windows first)
Lazy loading: delete compile.ts --no-bundle once a Deno version bump has shipped bundled
Swampify: @swamp/agent-sessions copies core's contracts, and assess writes candidate set v2
Collective invite shows "This invitation has expired or has already been used" after a successful join via Google sign-in
swamp help outputs in json instead of text
Respect dependency versions pinned with extension pull dep@version
serve: swamp serve reload turns scheduling and the timed services on or off and re-arms their intervals at once
serve: resuming a run waiting at an approval gate answers 500 action_failed on /api/v1/action
workflow run: every save converts the whole run record to YAML with @std/yaml, the largest remaining cost of a long run
workflow run: every run-record save re-reads and parses the whole record just to find the previous status, so cost grows with step count
Parallel steps on one model instance leave its lock idle between turns: hand it over in process, with a fairness cap for other processes
Swampify: assess method in @swamp/agent-sessions mines transcripts into a candidate set
serve: swamp serve reload changes grant mode, grant sources and explicit-grant rules, and removes admins, without a restart
Upgrading an extension by name leaves its manifest dependencies at their old versions
Find a working-directory Swamp repo even with `SWAMP_SERVE_URL` in the environment
serve: a reload that closes admission locks the admins out, and a boot-only continuation sweep reports a changed interval as applied
Lazy loading L6: opt-in bundled build (compile.ts --bundle)
serve shutdown: /ready reports shutting_down while the webhook route still queues new deliveries
Open serve links to every collective on an active trial or paid plan, replacing the built-in list and SERVE_LINKS_COLLECTIVES
A serve set up through the normal OAuth path can't link people: its client isn't a collective's, and the app only sees club_unavailable
Lazy loading: route completion callbacks to the target command instead of the full tree
Test broken on main: SqliteReadModel.recover full_recovery_required case merges into a directory that does not exist (found by #4612)
Lazy loading L4: load only the invoked subcommand in behavior and model
Flaky property test: TrustedHosts generator produces xn-- labels that parse correctly rejects
serve: pass swamp-club's links_require_plan refusal to the app, and record links ended by a lapsed plan
Team mode: a member's recovery after an interrupted command replays the whole history (minutes at a day of history); recover only the affected data names (starts after #4579)
Lazy loading L3: load only the invoked subcommand in data and workflow
serve: swamp serve reload applies the access lists, audit, intervals and proxy settings without a restart
serve: accept a link's return code on redeem and forward it to swamp-club
Lazy loading L5: one anchor for embedded resources
Lazy loading L2: route table, load only the invoked top-level command
Swampify: candidate set and comparison contracts in @swamp/agent-sessions
@swamp/kubernetes: all methods fail when kubeconfig tls-server-name is an IP address (kubectl accepts it)
serve: send instanceId in the swamp.club heartbeat body
Flaky: serve_team_write_round_trips_test checkpoint-interval request pinning sees an extra partition commit and chunks/has
fan-out: a failed run can end with its canary gate still pending, so a plain retry is refused with a hint naming the gate
fleet fan-out gates property test fails: retry after retryFrom diverges from the model (seed 1290163922)
Two source files carry a truncated licence header, and the header check only looks at the first line
deno run compile leaves the version stamp in the tree when deno compile fails
Lazy loading L1: move VERSION and GIT_SHA out of the version command
serve: reload installs trigger overrides from serve.yaml without validating them
swamp serve reload applies every setting that can change in place, validated and reported (auth lists, audit, limits; no --hot-reload needed)
datastore setup extension skips ${{ }} config interpolation that the load path applies — no clean setup for config-only-credential datastores
fan-out: a halt during a --from retry can name a lane as both failed and held, and the run-record judge refuses the save
test(workflows): extend the fan-out gates model test to lanes with more than one step
Scheduled flaky-test detection is failing
Swamp Club collective GUI: show the swamp serve instances currently connected to a collective
datastore setup extension over serve: resolve --config expressions with per-caller authorization
Team mode: serve-driven and member runs of one behavior collide on its method summary ('version 1 is already live'), and a report lease lapse can fail a run whose steps succeeded (#4567 follow-ups)
Lazy loading L0: startup benchmark and CLI parity scripts
Docs: serve now ends links on its own after a leaked client secret
Delegation: keep serve's links in step with swamp-club after linking
Team mode: a member's recovery after an interrupted command fails with 'Cannot apply commit N of partition run--…' (from #4374 run 5)
Scoped push queues each key twice: every data file is uploaded twice and every delete is sent twice
Audit OAuth client deletion
Collective-registered OAuth clients can't complete authorization_code: redirect URIs are stored as a JSON array string
Telemetry still adds about 0.9 s to every command on the dogfood VMs despite the detached sender (F33, follows #4511)
Team mode speed: far-member workflows within 1.5x v2, fewer commits per step and workflow, and no slow fresh clone, cold read or teardown after large outputs (F41, F42 from #4374 run 6)
Team mode: after a serve failover a member's writes stall 1-7 minutes, and takeover takes ~8 minutes under the default KillMode (F24, F14 from #4374)
swamp serve HA: a pending-run sweep can replay a queued run twice across hosts (heartbeat read error counts as a dead owner; control-plane-only runs have no claim compare-and-swap)
swamp serve: restarts without failed requests (close the listener early, overlap restarts, clients and workers ride out a restart)
Docs: step concurrency caps only that step's forEach iterations, not its whole dependency level
Docs: schedule catch-up and missed-fire tracking for scheduled workflows
Docs: what serve instances must not share, flat AWS Secrets Manager names, and where OAuth values are stored
Docs: usage-telemetry reference should describe the telemetry stats empty state and the last-flush record
Docs: Claude audit hook covers PowerShell, swamp audit install --user, and doctor audit hook warnings
Manual: the REST API page still says cancelling a workflow run needs admin
Docs: swamp dashboard --port, --trusted-hosts and its new lifetime (swamp-club#4513, #4509)
Docs: webhook payload is only in scope in trigger.inputs; document run id header and route reload
Docs: serve skips webhook routes under its reserved path roots; recommend /hooks/
Docs: webhook routes from reviewed route files (webhook-routes/)
autoDefinitions retention follow-ups: release failure misreports removals, step wording, quoted log list, lanes doc
software-factory: a work item whose slug extends another's reads the other's records
Team mode: a member's workflow run fails when another member's overlapping run commits the shared workflow report version first ('version N is already live'), about 1 run in 8 under load (F40 from #4374)
Retention or GC for auto-definitions whose runs and data are gone
serve: on macOS a webhook routes directory that does not exist yet under extensions/ escapes the 'holds extension code' refusal
serve: optionally poll the webhook route directory so a deploy alone makes routes live
feat(workflows): fan-out approval gates: rollout.approve suspends a fan-out job before dispatch with a prompt built from the frozen plan, and after the canary batch
serve: a run that reads suspended is refused as busy by approve/reject, and the busy reply hides ambiguous_gate on POST /api/v1/action
serve: reload installs webhook entries from serve.yaml without validating them
serve: a vault outage during swamp serve reload removes every working webhook route file
@swamp/gcs-datastore (and @swamp/s3-datastore): a 16 MiB chunk put times out after a flat 30 s, so a slow serve-to-GCS link can never store it
Team mode: a large output (64 MiB) takes about 8 minutes to store and usually loses its lease mid-upload, losing the output (F35 from #4374)
serve: load webhook routes from reviewed files in the repo
Team mode: a commit's cost on serve grows with history
Skill workflow reference uses range() in a forEach example; swamp CEL has no range(), and workflow validate passes it anyway
Slackwater: the reference Slack app, linking people and acting for them through serve's HTTP routes
worker connect: a relative --cache-dir enrolls fine but every dispatch fails with 'Path must be absolute'
serve: a webhook route can shadow serve's own API paths; require a fixed route prefix
swamp-club: move swamp-lib from the archived @systeminit scope to @swamp-club/swamp-lib
swamp-club: move from the retired @systeminit/swamp-lib to @swamp-club/swamp-lib
swamp-lib: link a person, act over HTTP and refresh delegated tokens
swamp-club: enable serve links for skunk-ape-test without an environment change
API runs out of Mongo connections under bursts: extension push timeouts, membership failures and UAT 500s
Ambient tags: set tags once per repo/worktree (env or .swamp), add them to every run and everything downstream, filter run history by tag
Allow editing recent feed submissions
Team mode: a member write takes 2 requests to serve instead of 8, workflows batch per step, and write cost stays flat over a day of history (performance from #4374)
feat(workflows): fan-out rollout: lanes start in canary batches under an error budget (maxFailures as a count or a percentage), hold the rest on a halt, and resume continues the rollout
Delegation: link a person to an integration and issue their tokens
Split the Litestream restore test file so it no longer sets the test suite's wall time
swamp-lib: workflowRun never settles when a run suspends at an approval gate
Run the test suite in several smaller processes, so verification is faster with no test weakened
swamp-lib: workflowRunStream drops a refusal and ends as if nothing went wrong
Generated @swamp/aws StateSchemas mark conditional and create-only properties as required, producing schema warnings on every read
fix(datastore): the run-record judge freezes a fan-out job's snapshot with its plan, so a forwarded record cannot rewrite the selector or worker scopes a resume trusts
Serve restart after kill -9 leaves the killed run reading running; only CLI recover's dead-owner check settles it
control_plane_schedule_state_repository_test: 'list skips a record whose read times out' fails under load because the 20 ms read bound also covers the good record
Scheduled flaky-test detection is failing
Cloudflare codegen: a nested properties-beside-oneOf required list can be read as the PUT body's complete required list
Team mode: member writes fail with GCS 429 once a team commits faster than once a second, because serve rewrites the lineage head object on every commit (F38 from #4374)
No way to add or switch to password login on an account created with SSO
Native interactive report views in the dashboard, with scoped workflow refresh
A member's behavior run rewrote its read model for 19 minutes (6.9 GB) and ignored SIGTERM (F37 from #4374)
feat(workflows): fan-out runs resume, retry and recover on their frozen plan, re-running only the lanes that did not succeed, and show the plan in log and JSON output
Observability never slows or breaks a command: the span exporter caps its connections, and no command waits on telemetry or pays slower start-up (F36, F33 from #4374)
Codegen/AWS: create-only required fields block list/get/delete in generated AWS models
Cloudflare codegen: a self-referencing request body overflows the stack in normalizeProperty
Cloudflare codegen: a oneOf/anyOf inside a multi-member allOf is dropped
swamp dashboard: trusted-host / bind option so it can be reached through a reverse proxy
issue-lifecycle: code conformance review is the final gate before the PR, after verification
A worker enrolling into a waiting queue can get its step before it can accept it, failing the step
feat(workflows): the fan-out lane engine: a fanOut job plans against serve's worker pool and runs one pinned lane per matching worker
deployment: getLogs method (deployment-aware pod log tailing)
Add expiring service-account token requests and authorization checks to RBAC
Agents.md: consider not using H1 'Project' inside managed block
Scheduled flaky-test detection is failing
Terminal report renderer prints HTML entities in table cells: a ' in a header shows as '
Slackwater: integration registration and ceiling
Rename username williamjtlacroix -> liywjl
Telemetry integration tests can fail verify-build by leaving a temp directory behind, and try to send telemetry to production
Slackwater: swamp-club link page, consent and the link email
verify-build can hang for up to an hour when test workers stop loading, and type checks during tests add heavy CPU load
swamp workflow validate fails a repo because of a pulled extension's bundled workflow (@keeb/setup-docker resolves fleet against the repo's own behavior)
Expose a manual_approval gate's decision identity (decidedBy) to workflow expressions
data query rejects CEL list macros: exists(c, ...) fails with Unknown field "c"
workflow approvals offers a person refused on their own run a cancel that serve refuses without admin
Over-limit contentMetadata publishes a blank extension page; archive fallback extracts nothing
workflow run --json reports a succeeded run as failed, and history get --json crashes, when resource content is nested past about 5,250 levels
Slackwater: bearer-only HTTP route for workflow actions
Deeply nested resource content overflows the stack: nested workflow run summaries are silently erased, and history outputs --json leaves truncated JSON
A method run writing thousands of resources reports 'Invalid string length' after succeeding, and runs out of heap past ~4,200
repo prune follow-ups: report failed writes apart from skips, resolved path in errors, upgrade note indent, matcher tests
Slackwater: grant conditions can read the client and the run's initiator
Slackwater: spike on delegated credentials through serve
repo upgrade tells a repo with tools: [] from an older upgrade to prune claude hooks it still uses
telemetry_workflow_method_invocations_test can leave its swamp-tel-wf-test temp dir behind, failing verify-build's leftover check
run summary box: tabs, carriage returns and non-color escape codes in a value still misalign the border and reach the terminal
repo upgrade: drop stale Claude audit hooks when claude leaves the tool set
behavior/workflow run: summary box border misaligns on rows with a truncated value
A workflow with a 500-step forEach takes 7m22s to complete after its last step has finished
A workflow run keeps every step's resource contents in memory, so a large forEach fan-out can exhaust the heap
Image factory: golden GCE and Docker base images built on ops
Ops: package the dossier workflows; give access:* a git path
Slackwater: record the client on requests, audit events and run decisions
Team mode: a member write makes one catch-up, its spans join serve's trace, and CLI start-up, fresh clones and cold reads stop costing seconds (group 4 latency from #4374, follows #4401)
Team mode: an admin can reveal a rotated token through serve, and lease, lock, grant and run-attribution answers are clear for members (F29, F9, F10, F8, F3, F27, F4 from #4374)
runModel infers incomplete extension owner for nested GCP package types
workflow run: show artifact attributes for runs through serve, and let a run leave out the "Data produced" box
Dashboard cannot show a step's log once the step's behavior definition is deleted, though the log data remains
@swamp/cve/sshole follow-ups: backport detection, file-read access rules, sshd.pam, Include, docker run timeout
Team mode: a member's command ends honestly when serve refuses, loses its lease or is unreachable (exit contract, F18, F7, F2, F26, F15 from #4374)
Workflow-scoped data get reports the behavior id as modelName once the behavior definition is deleted
Dashboard data item page cannot show an item once its behavior definition is deleted, though the data remains
@swamp/cve/tailworm: fixedVersion=1.104.4 would clear conn25 and funnel findings Tailscale did not fix; track fixes per finding
@swamp/cve/tailworm: follow-ups from #4456 review (TS_SSH_DISABLE_FORWARDING edge cases, raised fixedVersion text)
Drive Files: binary media upload and download support for Shared Drives
/cve/tailworm says no fix exists, but Tailscale fixed findings 1-3, 5, 7, 8 and 11 in 1.104.4 (TS-2026-014 to 017)
Add @swamp/cve/sshole: read-only detector for the OpenSSH musl password bypass
Slackwater: let customers build integrations that drive workflows securely and auditably
workflow run: let a run skip the "Data produced" box, or make it show what each artifact holds
@swamp/cve/emudhra: Ubuntu's pip vendored certifi reported as exposed although pip uses the OS store
@swamp/cve http2shell, php-rce-2026-10 and tailworm: scan fails with 'Duplicate data instance name' (log and status share a name)
stagecraft: rename work mode `method` to `behavior`, matching swamp's rename of models to behaviors
Container run inspection and isolation metadata for audited recovery tests
Team mode: serve confirms replication before a graceful stop, confirms each commit's own txid, and accepts tokens once ready (F22, F19, F30 from #4374)
Claude audit hook misses PowerShell tool calls and sessions started above the swamp repo
Flaky: serve_team_write_round_trips_test two-step run request pinning sees an extra commits read
Cloudflare codegen: disjunction branches that disagree on a scalar property type keep only the first branch's type
@swamp/video: informal demos — join multiple takes into one video, and a full-length background music bed
@swamp/video: small follow-ups from joining takes
@swamp/video: rendering an edit list that keeps nothing fails with 'sumExpr needs at least one term'
@swamp/video: splice a take into another at a point in edit.txt
Relationships
#4456 @swamp/cve/tailworm: fixedVersion=1.104.4 would clear conn25 and funnel findings Tailscale did not fix; track fixes per finding
Opened by stack72 · 11/5/2026· Shipped 11/5/2026
Tailscale 1.104.4 fixes Tailworm findings 1-3, 5, 7, 8 and 11 (TS-2026-014, 015, 016, 017, published 2026-10-26). Findings 4 (conn25 app connectors), 6 (headscale), 9 (funnel) and 10 (TS_SERVE_ALLOW_ALL_INTERFACES) remain unfixed.
In @swamp/cve/tailworm 2026.11.05.1, FIXED_TAILSCALE_VERSION is null and the fixedVersion global argument sets one patched flag that suppresses every code exposure (SSH, serve localapi, serve unix, app connector and funnel). Setting fixedVersion=1.104.4, as the scan output's own recommendation says to, would report a node running an app connector or funnel on 1.104.4 as clean. The extension also tracks no Tailworm bulletins, does not check finding 11, still prints No fixed Tailscale release yet, and mitigate has no option for the vendor workaround TS_SSH_DISABLE_FORWARDING=1.
Fix: record the fix version per finding (1.104.4 for 1-3, 5, 7, 8 and 11, none for 4, 9 and 10), add TS-2026-014 to 017, update the recommendations and manifest description, add a mitigate option that sets TS_SSH_DISABLE_FORWARDING=1 on tailscaled, and publish. Then refresh the sample outputs on /cve/tailworm in swamp-club from a real run of the new version.
Shipped
Click a lifecycle step above to view its details.