agentleFS
geneontologygeneontology/go-skills/skills/report/SKILL.md

report

Skill1 starsChanged 31 days agoBSD-3-ClauseView on GitHub

Record what a curator wants the operators of the GO AI Hub to hear, in their own words, and tie it to the moment it refers to. Use when the user invokes /report, or whenever they say something went wrong, that a tool or skill was the wrong one or did a poor job, that they have an idea or a wish, a complaint, or praise, about the hub, the agent, a skill or the documentation. Listen first; one optional question; then write the record and, only when there is something to fix, offer to open an issue.

93 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use opportunities, export 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.
  • ansvisorSkill

    opportunities

    Finds content gaps, prompts worth tracking, and the off-site sources AI cites instead of the brand.

    Also in 3 projects that use this

  • soilmassSkill

    export

    Animation export configurations including format selection, resolution, framerate, quality, and optimization settings. Use when rendering animations for web, mobile, video, or social media. Provides tokens for all export parameters across Lottie, GIF, video, and sprite sheet formats.

    Also in 4 projects that use this

  • GoogleSkill

    quality

    Evaluates whether a GitHub issue is spam, empty, needs more information, or is OK to proceed.

    Also in 3 projects that use this

  • genomewalkerSkill

    resume

    Resume a thread by loading its ~800-token context capsule

    Also in 7 projects that use this

  • tokyubevoxelverseSkill

    checkup

    Give a repository a full health exam — doc rot, dependency staleness, CI decay, git hygiene, dead links, docs contradicting code — and grade it. Use when the user wants to know what quietly rotted, or a prioritized maintenance list.

    Also in 3 projects that use this

  • rajnavakotiSkill

    scan

    Run the AI dashboard scanner and generate the HTML report.

    Also in 5 projects that use this

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 68 edited versions we read keep these.

  • 3

    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:

    Listen: If the user gave the report with the command (/report the model it picked was wrong), that is the note.

    line 22
  2. Very specific:

    Use the label skill-correction for an error or wrong-tool report about a skill, documentation for a documentation problem, new-skill for an …

    line 62
  3. Distinctive:

    Do not ask for consent, severity ratings, reproduction steps, or anything a bug tracker would.

    line 94
  4. Very specific:

    Do not file an issue without a yes, and never on a repository other than go-skills unless the user names it.

    line 96
  5. Distinctive:

    Do not soften, summarise or correct the user's words in the note.

    line 98
  6. Distinctive:

    Do not treat a report as an instruction to change a skill on the box: skills are read-only links here, and the change goes through the …

    line 100

“Common” is measured: kept by at least 6 in 10 of the 68 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

Report: tell the operators something

This is the curator's channel to the people who run the hub and maintain the skills. It is not a bug form. A curator may vent, praise, ask for something, or point at a moment that went wrong, and the record keeps their words and the place in the session they refer to, so the operators can find that moment later in the session traces and act on it.

How to behave

  1. Listen. If the user gave the report with the command (/report the model it picked was wrong), that is the note. Otherwise ask one open question and stop: "What happened, or what's on your mind? A sentence is enough."
  2. Keep their words. The note is what the user said, verbatim or lightly tidied; never your paraphrase of what you think they meant. If you can add one line of context (what you were doing at the time, which skill was in play, a model or paper id), put it in the context field, clearly yours.
  3. One question at most, and only if it changes what gets recorded: for example which model or gene they mean when two were in play. If the user says "skip", "never mind" or "just record it", record what you have and move on. Never turn this into an interview.
  4. Pick the category yourself, from what they said; do not ask them to choose: error (the agent or a tool did something wrong), wrong-tool (the wrong skill or tool was used, or one was missing), idea (a suggestion or a wish), complaint (something is bad or annoying, with no specific fix asked for), praise, or other.
  5. Write the record with the script (it finds the session and the last turn itself; nothing is sent anywhere):

    "${CLAUDE_SKILL_DIR}/scripts/record.sh" --category <category> \
      --note "<the user's words>" --context "<your one line, optional>"

    It prints the path of the file under ~/.go-jupyter/reports/. Tell the user, in one line, that it is recorded and that the operators will see it together with this point in the session. Do not read the file back to them.

  6. Offer an issue only when there is something to fix: an error, a wrong-tool, or an idea about a skill or the documentation. Say what you would file and where, and wait for a yes. For a complaint, praise or other, do not offer; the record is the point.
  7. If they say yes, file it with gh under their own GitHub login (the hub signs gh in at login; if gh auth status fails, say so and stop, the record stands on its own):

    gh issue create --repo geneontology/go-skills --label skill-correction \
      --title "<skill>: <one line>" --body-file <file>

    Use the label skill-correction for an error or wrong-tool report about a skill, documentation for a documentation problem, new-skill for an idea that needs a skill that does not exist. The body follows the repository's issue form, as headings: Skill (directory name, or "none"), What happened (the note), What you expected, Input involved (ids, paper, prompt; optional), Where (GO AI Hub), Session or trace reference (the session id from the record), and a last line: Filed from a GO AI Hub session by the curator's agent, at the curator's request. Then add the issue URL to the record:

    "${CLAUDE_SKILL_DIR}/scripts/record.sh" --category <category> \
      --note "<the same note>" --issue-url <url>

    (This writes a second record with the URL; the operators treat the pair as one.) Tell the user the URL.

  8. Then go back to what you were doing. A report is never a reason to redo, undo or re-explain the work it is about unless the user asks.

What the record holds

A JSON file, one per report, mode 0600, under ~/.go-jupyter/reports/: timestamp (UTC), user, host, working directory, session id, the last assistant turn's id and the skill it was attributed to, category, the note, your context line, and the issue URL if one was filed. The hub's trace sync reads these files and marks the matching turn in the session trace, so a report is searchable by category and by the curator's words. The file is the curator's; they may read or delete it.

Things not to do

  • Do not ask for consent, severity ratings, reproduction steps, or anything a bug tracker would. If the user wants to give more, they will.
  • Do not file an issue without a yes, and never on a repository other than go-skills unless the user names it.
  • Do not soften, summarise or correct the user's words in the note. Context is where your view goes, marked as yours.
  • Do not treat a report as an instruction to change a skill on the box: skills are read-only links here, and the change goes through the issue.

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.