Skip to main content

Obsidian Task Scan

@puddleworks/obsidian-task-scanv2026.11.10.1· MIT· 11d agoBEHAVIORSWORKFLOWSACCEPTED WARNINGS
01README

Find the open tasks in an Obsidian vault that a swamp extension could do, using Jev to classify them, and log each one no extension covers in Capture.

02Behaviors1
@puddleworks/obsidian-task-scanv2026.11.10.1obsidian_task_scan.ts

Global Arguments

ArgumentTypeDescription
vaultPathstringAbsolute path to the Obsidian vault folder.
excludeFolders?arrayVault folders to leave out, relative to the vault root. Daily notes (YYYY-MM-DD.md) are always left out.
batchSize?numberMost questions to send Jev in one request. Jev's free tier turns away requests of about 800 questions.
logAs?unionWho the Capture requests are logged for; they must be in Capture's people list. Empty logs nothing: the run is a preview.
captureModel?stringName of the @puddleworks/capture/requests behavior to log requests with.
captureUrl?urlWhere Capture's request pages are served, with no trailing slash; the result links to <captureUrl>/requests/<id>.
matchThreshold?numberLeast confidence in Jev's pick of an extension for the candidate to count as covered, and not be logged.
fn run(folder?: string)
List the open task lines ("- [ ]") in an Obsidian vault, leaving out templates and daily notes, and prepare a question for Jev about each.
ArgumentTypeDescription
folder?stringSubfolder of the vault to scan, relative to the vault root. Empty scans the whole vault.
fn classify(tasks: array, answers?: array, threshold?: number)
Classify each task from Jev\'s answers: a candidate when the answer is "extension" with at least the threshold confidence.
ArgumentTypeDescription
tasksarrayThe items from `run`.
answers?arrayJev's answers for each request that succeeded, keyed by task id. Tasks in a request that failed stay unclassified.
threshold?numberLeast confidence in an "extension" answer for the task to be a candidate. Jev\
fn match(tasks: array)
Search the swamp extension registry for each candidate, and prepare a question for Jev about which extension found, if any, already does it.
ArgumentTypeDescription
tasksarrayThe items from `classify`.
fn log(tasks: array, matched?: array, answers?: array)
Log in Capture each candidate no extension covers, as a draft request, and store every task's outcome. With logAs empty it only previews.
ArgumentTypeDescription
tasksarrayThe items from `classify`.
matched?arrayThe items from `match`.
answers?arrayJev's pick answers for each request that succeeded, keyed by task id.
fn candidates()
Answer which open tasks a swamp extension could do, from the latest scan. It changes nothing outside its own answer, and doesn't start a scan.
03Workflows2
@puddleworks/obsidian-task-scan0f449c58-6be3-4b5f-a3f4-5c106f6fd37c

Find the open tasks in an Obsidian vault that a swamp extension could do, using Jev to classify them, and log each one no extension covers in Capture.

scanList the vault's open tasks, leaving out templates and daily notes, with the questions for Jev in batches.
1.runobsidian-task-scan-model.run— The open tasks and the questions for Jev about them.
askAsk Jev whether a swamp extension could do each task, one request per batch. A failed request leaves its tasks unclassified.
1.ask-${{ self._index }}jev.ask— Ask Jev about one batch of tasks.
classifyClassify every task from the answers Jev gave.
1.classifyobsidian-task-scan-model.classify— Every task with its classification (candidate, not_candidate, unclassified or withheld) and confidence.
matchSearch the extension registry for each candidate, with a question for Jev about which extension found already does it.
1.matchobsidian-task-scan-model.match— What the registry search found for each candidate, and the questions for Jev's pick.
pickAsk Jev which extension found, if any, already does each candidate, one request per batch. A failed request leaves its candidates unlogged until the next run.
1.pick-${{ self._index }}jev.ask— Ask Jev about one batch of candidates.
logLog each candidate no extension covers in Capture as a draft request, and store every task's outcome. With logAs empty it only previews.
1.logobsidian-task-scan-model.log— Every task with its classification, confidence and outcome: logged (with the Capture link), matched, skipped or failed.
@puddleworks/obsidian-task-scan-candidatesebb91017-6baf-42db-b3da-cbd62c4c37eb

Say which open tasks in the vault a swamp extension could do, from the latest scan. The puddlerouterRecipe tag is the recipe puddlerouter reads: the phrase it matches and the data to show. It calls the same method as this workflow's one step, so running this workflow by hand gives the same answer. It starts no scan.

mainBuild the answer from the latest scan.
1.candidatesobsidian-task-scan-model.candidates— The candidates from the latest scan, and when it ran.
04Previous Versions5
2026.11.05.1MIT

Modified 1 behaviors. Added 1 workflows. updated labels

2026.11.02.2MIT
2026.11.02.1MIT

Modified 1 behaviors. Modified 1 workflows. updated dependencies

2026.10.30.1MIT

Modified 1 workflows

2026.10.27.1MIT
05Stats
A
100 / 100
Downloads
19
Archive size
27.6 KB
  • Has README or module doc2/2earned
  • README has a code example1/1earned
  • README is substantive1/1earned
  • Most symbols documented1/1earned
  • No slow types (deprecated)1/1earned
  • Dependencies pass trust audit2/2earned
  • Docs keep secrets in a vault1/1earned
  • Has description1/1earned
  • Platform support declared (or universal)2/2earned
  • License2/2earned
  • Verified public repository2/2earned

07Platforms
08Labels