agentleFS
AkkuAkku/stem/.claude/skills/port-check/SKILL.md

port-check

Skill1 starsChanged 15 days agoMITView on GitHub

Adapt STEM to an environment that differs from the reference (other OS, agent or model), measure whether it works, and report back. Use when session start prints "PORT CHECK AVAILABLE" and the user agrees, or when the user asks to run STEM on another OS, agent or model.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

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

    Missing prerequisite → abort: Write PORT.md (status: aborted), name the hurdle and what the user could do (for example Git Bash or WSL2 on …

    line 12
  2. Very specific:

    If a step did not fit: Learning: skill port-check: <what> -> <change> in the task notes (R-LEARNING).

    line 18

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Port check

Reference environment: see reference-* in STEM_VERSION. Checklist, failure categories and report format: docs/agent/PORT_CHECK.md (read it first). STEM files stay untouched: adaptations go into the overlay docs/agent/port/ and the log docs/agent/PORT.md, so stem-update does not overwrite them.

  1. Offer once. One question to the user: run the port check now (a/b)? On no, create docs/agent/PORT.md with status: declined and the environment (stops the offer), then continue the normal task.
  2. Analyse (read only): confirm OS, agent and model with the user (the script knows only OS and agent; compare your own model with reference-model). Run the stage 1 prerequisite probe (PORT_CHECK.md): bash, python3, git, backlog CLI, hooks support.
  3. Missing prerequisite → abort. Write PORT.md (status: aborted), name the hurdle and what the user could do (for example Git Bash or WSL2 on Windows), then go to step 7. Substitutes (stage 2) only if the user then chooses that.
  4. Plan. Show one plan: which scripts or instructions need an adaptation, which overlay file each goes to, what you expect to fail. One approval for the whole plan, then no further questions on it.
  5. Apply as overlay: adapted copies in docs/agent/port/, deviations and environment in PORT.md. Do not edit STEM files.
  6. Measure. Run the stage 1 checklist, record per item pass / partial / fail with failure category, the user action a failure would need (and whether the user did it) and the number of user interactions per flow. Stage 2 (optional, only if all pass): scripts/retro-metrics.sh --snapshot, compared with other baselines as order of magnitude only (agent and model are confounders).
  7. Report. Build the report from PORT.md (numbers, IDs and categories only; format in PORT_CHECK.md), show it to the user in full, send only on an explicit yes. Adaptations that pass the checklist may go to STEM with promote-learning; the user decides.

If a step did not fit: Learning: skill port-check: <what> -> <change> in the task notes (R-LEARNING).

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.