agentleFS
MoeenNehzatiMoeenNehzati/famulus/skills/using-compass/SKILL.md

using-compass

Skill4 starsChanged 4 months agoMITView on GitHub

Use when a user or another skill directs the agent to use a named compass.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 0 are distinctive, and 3 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.

  • 3

    Very specific

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

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

    Use one invoker-provided authorized VoyageDispenser process binding supplied with the activation request.

    line 18
  2. Very specific:

    Use only the IDs returned by initiate.

    line 25
  3. Very specific:

    After reading and retaining a terminal result, it invokes release for its assigned Voyage unless there is an explicit reason to preserve …

    line 39

≈ “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
<!-- BEGIN BLUEPRINT INTERFACES -->

Generated from blueprint.yaml. Do not edit this block by hand.

Instruction Interfaces:

These are LLM-readable instruction surfaces. Read and follow them directly; do not invoke the MCP server for them.

  • rutter.interface.dispenser@5 — Construct one VoyageDispenser from self-describing modes, unique run initialization, optionally prefix-filtered ID enumeration, Voyage resolution, and Voyage-or-run release callbacks.
<!-- END BLUEPRINT INTERFACES -->

Using Compass

Use compass on <rutter-name>.

Use one invoker-provided authorized VoyageDispenser process binding supplied with the activation request. If it is absent, report a public-interface gap and stop.

First invoke help and follow the returned operating contract. If the activation supplies the advertised mode arguments, invoke modes, then initiate exactly once to create a fresh run, passing the optional run prefix only when supplied. Use only the IDs returned by initiate. If initialization inputs are absent, report a public-interface gap instead of inventing them. Act as the controller: assign exactly one agent to each returned voyage_id, giving it the dispenser binding and its assigned voyage_id. Agents must not share or switch Voyage IDs. Do not start any Voyage agent until every returned ID has an assigned agent. If an independent agent cannot be assigned to every Voyage, report a public-interface gap and stop.

Each Voyage agent invokes status only with its assigned ID. For a ready Message, it performs the instructions and payload, produces a response satisfying the optional response schema, invokes validate, and invokes advance only after validation succeeds. For ready automatic work, it invokes advance without a response. It reads a fresh status after every successful advance and stops on terminal, fault, uncertain, malformed, or unknown status. After reading and retaining a terminal result, it invokes release for its assigned Voyage unless there is an explicit reason to preserve that Voyage's working directory. It reports that reason when it preserves the directory and never releases a ready, faulted, uncertain, malformed, or unknown Voyage. The controller must wait for every Voyage agent to stop before finishing.

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.