Lab #2410: Extension catalog caches a datastore bundle with an empty type, so the repo fails with Unknown datastore type when addressed by its realpath

Summary

A repo whose datastore is @swamp/s3-datastore fails every datastore-touching command with:

Unknown datastore type "@swamp/s3-datastore" in .swamp.yaml. Available types: filesystem

when the repo is addressed by one path, and works when it is addressed by another path to the same directory. The extension catalog holds a datastore row with an empty type for the failing path.

Found while reproducing swamp-club#2405 on macOS, swamp 20260923.142649.0-sha.379bb5f0, @swamp/s3-datastore 2026.09.23.1.

Reproduction

On macOS /tmp is a symlink to /private/tmp.

mkdir /tmp/repro && cd /tmp/repro && swamp repo init
swamp datastore setup extension @swamp/s3-datastore --repo-dir /tmp/repro --namespace ns --config <s3/minio config> succeeds and auto-installs the extension.
Run any command with the shell cwd at the repo (the shell reports it as /private/tmp/repro), e.g. swamp model create command/shell m1 --json.

Observed: auto_resolve reports failed ... reason: already_installed for /private/tmp/repro/.swamp/pulled-extensions/@swamp/s3-datastore, then the command fails with the error above.

After that, from the same machine:

Invocation	Result
`--repo-dir /tmp/repro`	works
cwd `/tmp/repro`	works
`--repo-dir /private/tmp/repro`	Unknown datastore type
cwd `/private/tmp/repro`	Unknown datastore type
cwd `/private/tmp/repro` with `--repo-dir .`	Unknown datastore type

Evidence

.swamp/_extension_catalog.db, table bundle_types, holds exactly one datastore row:

source_path     = /private/tmp/repro/.swamp/pulled-extensions/@swamp/s3-datastore/datastores/s3.ts
type_normalized = (empty string)
kind            = datastore

The row is keyed by the realpath form and has no type, so a lookup through that path finds a cached entry that registers nothing. The symlinked form has no row and appears to fall through to a fresh scan.

Expected

The catalog never persists a datastore row with an empty type, and the same repo resolves identically regardless of which path (symlink or realpath) addresses it.

Notes

Not yet traced in code. Worth checking whether the empty row is written by the failed already_installed auto-resolve path, and whether source paths should be canonicalised (realpath) before they key the catalog.

Session-start 'swamp model search --json' in the AGENTS.md template is unfulfillable - wire it to a hook or drop it

Flaky property test: analyzeExpression dot/bracket agreement fails when the generated model name is a CEL keyword

Accept usernames for user subjects in grant files

Docs: enable-managed-config should say to re-run config migrate when its push fails

Docs: grant file reference shows a stale format and omits resources and subjects

Flaky property test: analyzeExpression hyphenated dot access fails when modelName generates a CEL keyword ("in")

extension quality and push crash with IsADirectory when additionalFiles lists a directory

extension push: a required global argument renders with a stray "" ("text": "string""")

Docs: extension publishing reference is stale on bare specifiers, quality output and fmt config (swamp-club#3022)

swamp-extensions: replace bare import specifiers with npm:/jsr: specifiers in first-party extensions

extension push: 'Requested visibility' line prints to stdout without the logger prefix, twice per run

A nested structural swamp that outlives the run that started it keeps skipping that run's lock while the holder writes

Show signal waits to remote clients and the dashboard

Settle expired signal waits from swamp serve

Resume signalled workflow runs automatically under swamp serve

extension push/quality: structured acceptances for agents, a clearer title than For next time, and an optional reason

extension push: the undecided private-entitlement verdict still says the registry may start a trial at publish

seed:collective cannot produce a superseded personal trial or a returned trial, so the UAT cannot cover them

Replayable state fragments: Git-reviewable desired state for resources changed by model methods

extension push: registry metadata and content hash label files relative to the repo dir, so a push with --extensions-dir records ../ paths

registry: store and show declared acceptances sent in contentMetadata.acceptances at confirm (swamp-club#3021)

extension push --dry-run: authentication verdict prints 'Signed in as .' when the credential is an API key with no username

serve --remote-only: step output from workers is never persisted to the run log (history logs always empty)

serve HA: worker enrollment on a replica that lacks the enrollment token's definition may create a second definition (unverified)

serve HA: pull a token's definition on an auth miss so a peer accepts a new token at once

extension push: version-drift check reports 'no previously published version found' when the registry call failed

Search and list output follows directory iteration order in the definition, workflow, vault and output repositories

extension push: accept attestation-backed adversarial-review evidence (lookup by extension and content hash), generated provenance, and carry the evidence to the registry at confirm (design: Lab #3023)

Workflow load errors print the raw Zod issue dump instead of a readable message

Decide a stricter naming rule for workflow step and job names

serve: run scheduled workflows concurrently (per-workflow serialization), and report queue delay

Show expired approval gates as expired, with a Cancel action in the dashboard

digitalocean codegen leaves an orphaned security_secret.ts model file after its endpoint left the spec

extension promote accepts a yanked version and reports it promoted

Docs: doctor install reports a stale or failing autoupdate scheduler

extension push output: double-quoted adjacent interpolations, repo-relative paths, and --json emits three JSON documents

auto-resolve: the Installing line prints the extension's entire multi-line description

s3-datastore/serve: readResource without a version returns an older version; the latest marker is rewritten backwards

Publish agent-facing projections of the manual: /llms.txt, .md pages, and llms-full.txt

Let agent-runner drive more agent CLIs (Kilo Code and others)

s3-datastore: every fast-path miss re-downloads every _index shard, so a busy serve pulls the whole index every poll (~230 GB/day S3 egress)

Feedback: Swamp as an approval-gated control plane for a small fleet

extension safety analyzer's Deno.Command( warning is a plain substring match

Tell timeouts apart from cancels in method-run records, and review the hidden 30 s fallback timer for step-called model methods

Run from a git worktree: definitions from the worktree, state from the shared repo

S3/GCS datastore: a push that fails after uploading drops its recorded deletes, so the retry brings deleted data back

Tests: bring the in-memory remote's default semantics up to @swamp/s3-datastore and @swamp/gcs-datastore 2026.10.01.1

Tell the pushing client when its extension contentMetadata fails validation

Extension datastore: query still returns an item after its delete is pulled

Cancel, reject and supersede of a parent run should settle its suspended nested child runs

Tracking: test baseline required before the datastore refactor (commit-log design)

Tests: serve pollers make a peer's writes, deletes and grants visible with a real catalog (before Phase 5)

datastore setup can overwrite an existing remote config tier when it moves an in-repo tier into an extension datastore

serve: define and supply the collective and owner grant condition variables

No first-class workflow primitive for spawning a permission-scoped agent session

A pulled extension can shadow a built-in type, and removing it leaves serve without the built-in

Decorative animations outside the Lab still repaint on the main thread

s3/gcs datastore: pullChanged overwrites dirty, unpushed cache files

Concurrent first runs in a fresh repo log 'Catalog migration to per-extension-aggregate-v3 failed (database is locked)'

Codegen/DigitalOcean: create-only required fields block list/get/delete in generated DigitalOcean models

Codegen/AWS: create-only required fields block list/get/delete in generated AWS models

End a collective's trial when it begins a paid subscription

A collective that cancels inside its 30-day trial window regains trial access to private extensions

s3-datastore: SWAMP_S3_REQUEST_TIMEOUT_MS does not appear to apply to ListObjectsV2 during pull

s3-datastore: S3Lock.acquire overshoots maxWaitMs by up to a full backoff interval

s3-datastore: a model-scoped pull still lists, walks and indexes the whole namespace

Run the web dashboard locally from the CLI without swamp serve

verify-reviews: adversarial review path guard is a hand-kept list, so new extensions silently skip it

issue-lifecycle skill: prepare-to-ship documents a fast_forward method that does not exist

S3 datastore rejects valid cache paths as traversal on Windows

swamp-club: unknown collective API token returns 422 instead of the documented 404

Partial extension catalog saves delete rows for sources mounted from outside the repo

extensions: cold-path catalog rebuild skips sources whose type is not a string literal, dropping them for one process lifetime

serve: rotate the external token-secrets key

serve: server token GC follow-ups (upgrade backlog holds the sync gate, not-found matching, owner lookup, UX)

serve: without a remote datastore the token GC is not serialized against token rotate/re-mint

Workflow evaluation rejects another templating system's ${{ }} text, so the #2491 pass-through never reaches workflow steps

@swamp/aws/certificatemanager/certificate: expose ACM's reported attributes (NotAfter, Type, RenewalEligibility, InUseBy) on read

worker prune: remote datastore keeps deleted worker records (no-path markDirty() skips deletions)

Orphaned data record survives a model type migration and is unreachable by data delete/versions/prune