agentleFS
opendatadiscoveryopendatadiscovery/odd-team/.claude/skills/triage/SKILL.md

triage

Skill0 starsChanged 2 months agoApache-2.0View on GitHub

Convert raw scanner findings into atomic work items in the backlog. Assigns priority, category, effort, and identifies file conflicts.

190 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use writing-beats, setup-matt-pocock-skills and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • LesaniSkill

    writing-beats

    Shape an article as a journey of beats, choose-your-own-adventure style. The user picks a starting beat from the raw material, you write only that beat, then offer options for where to pivot next, beat by beat, until the article reaches a natural end. Use when the user has raw material and wants to assemble it as a narrative rather than an argument.

    Also in 74 projects that use this

  • GoogleSkill

    setup-matt-pocock-skills

    Configure this repo for the engineering skills — set up its issue tracker, triage label vocabulary, and domain doc layout. Run once before first use of the other engineering skills.

    Also in 182 projects that use this

  • SethGammonSkill

    ascii-diagram

    Generate perfectly aligned ASCII diagrams — architecture, flow, sequence, box-and-arrow. Uses a programmatic character-grid approach so alignment is guaranteed by math, not token prediction. Includes post-render verification.

    Also in 4 projects that use this

  • simo97Skill

    migrate-to-shoehorn

    Migrate test files from `as` type assertions to @total-typescript/shoehorn. Use when user mentions shoehorn, wants to replace `as` in tests, or needs partial test data.

    Also in 74 projects that use this

  • LesaniSkill

    writing-shape

    Take a markdown file of raw material and shape it into an article through a conversational session — drafting candidate openings, growing the piece paragraph by paragraph, arguing about format (lists, tables, callouts, quotes) at each step. Use when the user has a pile of notes, fragments, or a rough draft and wants help turning it into something publishable.

    Also in 70 projects that use this

  • liufeng1976Skill

    setup-ts-deep-modules

    Wire dependency-cruiser into a TypeScript repo so each package is a deep module — implementation hidden in subfolders, reachable only through its entry-point files. User-invoked.

    Also in 47 projects that use this

  • +1 more from Lesani

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 4 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 179 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Check existing items in backlog/ to avoid duplicates and get next ID numbers

    line 18
  2. Very specific:

    Follow format exactly from backlog/README.md

    line 43
  3. Distinctive:

    Write testable acceptance criteria (not vague goals)

    line 44
  4. Distinctive:

    Update backlog counts by category and status

    line 53
  5. Very specific:

    If $ARGUMENTS is empty, list available findings in findings/ and ask which to triage

    line 73
  6. Distinctive:

    Don't write implementation details — just acceptance criteria

    line 74
  7. Distinctive:

    If two items touch the same file, note dependency explicitly in both

    line 76

“Common” is measured: kept by at least 6 in 10 of the 179 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Triage Findings

Convert findings at $ARGUMENTS into backlog work items.

Protocol

  1. Orient — Read:
    • CLAUDE.md
    • backlog/README.md (work item format and lifecycle)
    • state/file-registry.yaml (existing file claims)
    • Check existing items in backlog/ to avoid duplicates and get next ID numbers
  2. Read findings — Load the findings file at $ARGUMENTS

2.5. Look up the scanner-feed log (rev 13):

  • Derive the scan_run_id from the findings file path. The matching scanner-feed log lives at lineage/{substrate_repo}/scanner-feed/{YYYY-MM-DD}-{scan_run_id}.yaml.
  • If the log exists (mode-B scan-run produced it): load it. The log carries clues_consumed[], agent_consultations, write_backs, warnings, and the per-finding findings_produced cross-reference.
  • If no log exists: the findings came from a mode-A standalone scan; no ontology-corroboration data is available. Proceed with normal triage.
  • If verification_class: corroboration-only: the scan-run did NOT iterate the feature catalog independently (Rule 21 D13 violation). Flag this in the triage report and recommend a re-run; findings can still be triaged but their coverage is suspect.
  1. Triage each finding — For each, decide:
    • Actionable? — Reject false positives with brief explanation
    • Work we do here, or a handoff to upstream? — This is the first split:
      • Work we do here → backlog item under backlog/{cat}/{ID}.md (continue with the fields below).
      • Handoff to upstream (the fix lives in a repo we don't directly maintain in this workspace — Java in odd-platform, Python in odd-collectors, schema in the spec, Helm in charts, etc.) → instead of a backlog item, draft a GitHub issue at issues/{repo}/{PREFIX}-NNN.md via /log-issue {repo} "title" (see issues/README.md for prefixes) and FILE it, writing the returned URL back into the draft — since 2026-09-06 filing is the agent's action, not a hand-off. A single finding can produce both — e.g., a missing doc caveat (backlog) plus an upstream code fix (issue draft); link them both ways.
    • Category (for backlog items) — DOC (documentation), TST (test), NAV (navigation), SPC (specification)
    • Issue type (for issue drafts) — bug | feature | adjustment; bugs also need severity
    • Granularity — One work item / one issue draft per logical change (split complex findings, merge trivial ones)
    • Priority — critical (wrong info) > high (missing tests for fragile code) > medium (gaps) > low (cosmetic). Rev-13 ontology-corroboration lift: if the scanner-feed log lists this finding's findings_produced against an ontology clue source AND the clue is either feature-flow or doc-gap corroboration, lift the priority by ONE tier (medium → high, high → critical). Two independent signals (ontology AND scanner) flagging the same gap is published-mistake risk per CLAUDE.md's "wrong docs > missing docs" priority bar. Concept-source clues (non-canonical-term) default to medium; shoebox-source clues inherit the shoebox thread's severity.
    • Release routing (documentation items — adrs/drafts/release-train-doc-gating.md Decision 3) — does the change describe behaviour absent from the latest published odd-platform release? Mechanical test: find the introducing commit; git -C ../odd-platform tag --contains {sha} empty (or only unpublished tags) ⇒ release-gated → set milestone: from the introducing issue/PR's milestone (GET /repos/.../issues/{n} → .milestone.title); no open matching milestone → create the item blocked with an unresolved-milestone note (the maintainer creates/re-targets milestones). Released-truth corrections — including known-issue caveats for released bugs — stay un-gated and publish immediately at their normal priority. A release-gated item's urgency is "complete before the release gate", not "before the next session".
    • Affected files — Which files will be created/modified? (For issue drafts: cite upstream file:line in the body, not affected_files in this workspace.)
    • Dependencies — Does this item require another to complete first?
    • Effort — small (<30min), medium (30-90min), large (>90min) — for backlog items only; issue drafts have no in-workspace effort once the body is written.
  2. Create work items — Write to backlog/{category}/{ID}.md:
    • Follow format exactly from backlog/README.md
    • Write testable acceptance criteria (not vague goals)
    • Include scanner source and found date
    • For findings sourced from ontology clues (rev 13): include ontology_source: in the backlog item's Context section — e.g. ontology_source: feature-flow F-007 (Source: Ontology[F-007:hop-2] → Repo[AlertManagerController.java:24-31]). This back-links the work item to the F-NNN it came from, so when the doc PR ships, the maintainer can flip the F-NNN's doc_status: field from backlog → drafted → live.
    • For release-gated items: set milestone: "{version}" in the frontmatter and record the affected pages + expected post-merge URLs in the item — the release gate (playbooks/release-train-merge.md) consumes them.
  3. Update file registry — Add to state/file-registry.yaml:
    • Map each affected file to the new work item ID
  4. Update progress — Edit state/PROGRESS.md:
    • Update backlog counts by category and status
  5. Report — State:
    • Findings processed: N accepted, M rejected (with reasons)
    • Work items created: breakdown by category and priority
    • Any dependency chains identified

Splitting Heuristics

Split when:

  • Finding spans multiple repos
  • Finding requires both code and doc changes in different repos
  • Finding affects >5 files

Merge when:

  • Multiple findings are same root cause
  • Multiple findings need the same single-file change

Rules

  • If $ARGUMENTS is empty, list available findings in findings/ and ask which to triage
  • Don't write implementation details — just acceptance criteria
  • Functional bugs in upstream code → draft an issue at issues/{repo}/{PREFIX}-NNN.md via /log-issue (never just narrate "log as GitHub Issue" in a work item; that disappears the moment the conversation ends). If a doc caveat also ships now, pair the backlog item and the issue draft with two-way links.
  • If two items touch the same file, note dependency explicitly in both
  • A finding that turns into both a backlog item and an issue draft must reference each other in their Context / discovered_during: fields — the audit trail is two-way.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.