agentleFS
nasanasa/fprime/.github/skills/fprime-checklist-closeout/SKILL.md

fprime-checklist-closeout

Skill12k starsChanged 21 months agoApache-2.0View on GitHub

Close-out pass of the F Prime Component Development Checklist. Run after system testing / run-for-record: re-runs fprime-auto-checklist with --stage closeout, decides the Close-Out (section 5) review rows, re-decides anything STALE from earlier stages, and audits every tool-decided row (YES/NO/N/A/UNCHECKED) so the delivered checklist has no row nobody read. Keywords: F Prime, checklist, close-out, audit, hardware run, CI, fprime-auto-checklist.

  • Reads credentials
  • Installs packages

Trending

More talk lately.about the whole repository

Dec 2025: 0 mentionsJan 2026: 0 mentionsFeb 2026: 0 mentionsMar 2026: 1 mentionApr 2026: 0 mentionsMay 2026: 22 mentionsJun 2026: 3 mentionsJul 2026: 7 mentionsAug 2026: 2 mentionsSep 2026: 7 mentionsOct 2026: 11 mentionsNov 2026: 1 mention
Dec 2025Jun 2026Nov 2026
Mentions a month of the nasa/fprime repository that our weekly crawl found. There is no adoption data for this file yet.

Who's talking about nasa/fprime

The biggest 5 of 71 mentions linking to the repository that our crawl found. Each line is their own headline.

See all 71

Talked about on GitHub, YouTube, Hacker News and the web (71 times).

On GitHub · 50

45 more on GitHub

On YouTube · 13

8 more on YouTube

On the web · 2

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 2 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    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. Distinctive:

    Run against the previous checklist:

    line 41
  2. Very specific:

    Use --from decisions.json for the batch.

    line 73
  3. Very specific:

    YES: no note. The tool's own rule pointer (R8 pass, ` cpp/fprime/no-printf 0 findings ) is the only text a YES row carries.

    line 85
  4. Distinctive:

    If the process has not done it, say so.

    line 105
  5. Very specific:

    Never hand-edit the Markdown table.

    line 106

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “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

Checklist — close-out

Finishes the component's docs/checklist.md. At this stage every section is due and the tool additionally queues each row it decided itself for an audit: a reviewer reads the evidence behind the verdict and confirms or overrides it. The skill performs no development step: whether the component was built into a topology, run on hardware, or passed CI is answered from the artefacts the process left (docs/hardware-run.md, the CI status via GITHUB_TOKEN, deployment build caches), truthfully — "not done" stays UNCHECKED, it is never argued into a YES.

Procedure

  1. Install the tool from a clone (the clang-tidy plugin is built from the clone, it is not in the pip package); without the plugin every static-analysis row (31 of them) comes back UNCHECKED, so do not skip it:

    git clone https://github.com/LeStarch/fprime-auto-checklist && cd fprime-auto-checklist
    pip install -e .
    sudo apt-get install -y clang-tidy-15 llvm-15-dev libclang-15-dev   # LLVM 15 (plugin and clang-tidy must match)
    cmake -S native -B native/build -DClang_DIR=/usr/lib/llvm-15/lib/cmake/clang && cmake --build native/build
    cd -   # back to the project root; run the tool from there

    A prebuilt plugin is picked up from $FPRIME_TIDY_PLUGIN (--tidy-plugin), clang-tidy from $CLANG_TIDY (--clang-tidy).

  2. Run against the previous checklist:

    fprime-auto-checklist run --module <Path/To/Component> --analyzer clang-tidy \
        --stage closeout --previous <Path/To/Component>/docs/checklist.md \
        -o <Path/To/Component>/docs/checklist.md

    (Add --repo owner/name with GITHUB_TOKEN set for 5.5; --artifacts-dir for 4.1 when the process published unit-test artifacts.)

    The queue contains review rows (section 5 and anything left undecided), stale rows (earlier decisions whose section inputs changed) and audit rows — every tool-decided row without a current audit.

  3. Decide the review and stale rows as in the earlier stages (rule text is in the queue). 5.3 is the hardware run: YES only when docs/hardware-run.md records platform, deployment, commit, date, result and tester for the current commit (check C2 validates it); otherwise NO with "not run" — never infer a run from a build.
  4. Audit every audit row. For each, open the evidence the Notes cite (the clang-tidy/CodeQL finding and file:line, the check's detail, the SDD table, the coverage figure) and record:
    • decide --item <id> --status <same as tool> --note "<what you checked>" — recorded as Audited: confirmed;
    • decide --item <id> --status <different> --note "<why the tool is wrong>" — recorded as Audited: overridden; the note must point at the evidence (a false positive's file:line, the artefact the tool missed). An UNCHECKED row is confirmed as UNCHECKED only by stating which process step has not been performed (--status is not needed for that: leave it and list it in the report); override it to YES/NO only with an artefact in hand. Use --from decisions.json for the batch.
  5. Re-run step 1 and confirm Nothing to review at the closeout stage: zero REVIEW rows, every row carrying a decision or an audit stamp. Report the totals, the NO rows with notes, and the UNCHECKED rows with the process step each waits on. Commit docs/checklist.md and docs/checklist.json with the component.

Notes discipline

The rendered table is read by people; keep it sparse.

  • YES: no note. The tool's own rule pointer (R8 pass, `cpp/fprime/no-printf` 0 findings) is the only text a YES row carries; your --note (one clause, the ids you checked) is stored in the JSON for traceability and is not rendered.
  • NO: one line, what is wrong and where — file:line, requirement id or table — and the fix in a few words. No requirement-by-requirement walks, no restating the tool's finding, no "Confirmed:" prefixes.
  • N/A: one clause stating why the item does not apply.
  • UNCHECKED: name the process step that has not been performed.
  • Audits that confirm the tool's verdict are silent in the table; an override renders as one Audit (...) line from your --note.

Rules

  • docs/checklist.md stays beside docs/sdd.md; the SDD's last line is Component Development Checklist: [checklist.md](checklist.md) — restore it if an SDD edit dropped it.
  • An audit reads the evidence; it does not re-run the analysis to make a finding disappear. Fix the code or the document, re-run the tool, and the audit lapses by itself (the tool drops an audit whose verdict or section fingerprint changed).
  • Hardware, topology and CI rows are answered from records of what was done. If the process has not done it, say so.
  • Never hand-edit the Markdown table; decide re-renders it from the JSON.

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.