agentleFS
TidalPaladinTidalPaladin/skills/goal-mode/SKILL.md

goal-mode

Skill1 starsChanged 24 days agoApache-2.0View on GitHub

Manage the current task's native goal only when the user explicitly invokes $goal-mode.

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. 0 are common, 7 are distinctive, and 6 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 7

    Distinctive

    This author's own choices. Worth a look.

  • 6

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use native goal tools for a persistent, verifiable objective.

    line 8
  2. Distinctive:

    Do not invoke this skill implicitly.

    line 10
  3. Distinctive:

    Write one concise, outcome-oriented objective.

    line 16
  4. Distinctive:

    Keep the objective self-contained and no longer than 4,000 characters.

    line 22
  5. Very specific:

    Call get_goal before deciding whether to change goal state.

    line 34
  6. Very specific:

    When an unfinished goal exists, preserve it.

    line 35
  7. Very specific:

    When no goal exists and the current request has a clear unfinished outcome, call create_goal with the objective derived above.

    line 36
  8. Distinctive:

    Leave Goal Mode disabled for trivial or completed tasks.

    line 37
  9. Very specific:

    Call update_goal with status=complete only when the objective is achieved and no required work remains.

    line 39
  10. Very specific:

    Use status=blocked only after the same blocker recurs for at least three consecutive goal turns.

    line 40
  11. Distinctive:

    Do not use terminal states merely to disable Goal Mode or escape difficult work.

    line 44
  12. Distinctive:

    If the active goal does not match the requested work, do not end it solely to replace it.

    line 51
  13. Very specific:

    If get_goal, create_goal, or update_goal is unavailable, do not claim that goal state changed.

    line 55

“Common” is measured: kept by at least 6 in 10 of the 3 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

Goal Mode

Use native goal tools for a persistent, verifiable objective. Manual invocation grants permission for the current task to inspect, create, and truthfully end a goal. Do not invoke this skill implicitly.

This permission does not expand filesystem, network, approval, Git, or GitHub access. It does not override Plan Mode or any other execution restriction.

Describe the Goal

Write one concise, outcome-oriented objective. Include these elements when they apply:

  • Outcome: State the result or behavior to deliver, not only the activity to perform.
  • Constraints: Name required scope, compatibility needs, boundaries, tools, or approaches to avoid.
  • Verification: State the tests, measurements, review result, or other evidence that proves completion.

Keep the objective self-contained and no longer than 4,000 characters. Point to a file when supporting instructions are longer. Include a token budget only when the user explicitly requests one.

Example:

Update the CLI to emit stable JSON errors without changing text output, add
regression coverage for malformed input, and make the repository's formatting,
linting, type-checking, and test gates pass.

Manage Goal State

  1. Call get_goal before deciding whether to change goal state.
  2. When an unfinished goal exists, preserve it. Do not call create_goal to replace it.
  3. When no goal exists and the current request has a clear unfinished outcome, call create_goal with the objective derived above.
  4. Leave Goal Mode disabled for trivial or completed tasks. Ask only when a necessary objective is unclear.
  5. Continue the task under the active goal. Use get_goal when current status or remaining budget matters.
  6. Call update_goal with status=complete only when the objective is achieved and no required work remains.
  7. Use status=blocked only after the same blocker recurs for at least three consecutive goal turns. Meaningful progress must require user input or an external state change. A resumed blocked goal starts a fresh blocked audit.

Do not use terminal states merely to disable Goal Mode or escape difficult work. A nearly exhausted budget does not establish completion. Leave a nonterminal goal active.

For a new authorized objective, call create_goal only when no unfinished goal remains.

Handle Existing or Unavailable Goals

If the active goal does not match the requested work, do not end it solely to replace it. Explain the mismatch and let the user edit or clear the goal before creating another one.

The current programmatic interface supports inspecting, creating, and terminally updating goals. Pause, resume, edit, and clear remain user-controlled operations when corresponding native tools are unavailable. The user can use the Goal Mode progress controls or /goal pause, /goal resume, /goal edit, and /goal clear.

If get_goal, create_goal, or update_goal is unavailable, do not claim that goal state changed. State which operation could not be performed and provide the equivalent /goal action when one exists.

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.