agentleFS
my-chiefmindmy-chiefmind/ai-maestro/skills/tracker/SKILL.md

tracker

Skill6 starsChanged 48 days agoMITView on GitHub

Verify the current runtime's external MCP tracker connection, link a local ticket, and perform explicitly requested task or documentation updates.

7 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 4 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 4

    Distinctive

    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:

    Read the project's config.json tracker preferences as data.

    line 8
  2. Distinctive:

    Use only the configured services, MCP connection names, and destinations.

    line 10
  3. Very specific:

    If missing, run maestro tracker connect --runtime <runtime>.

    line 19
  4. Distinctive:

    Treat documentation updates separately from task updates.

    line 48
  5. Distinctive:

    After each write, confirm the result through the tool response or a read.

    line 50
  6. Distinctive:

    No background sync, broad imports, deletion, or automatic status propagation is implied.

    line 53

“Common” is measured: kept by at least 6 in 10 of the 5 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

External tracker

Read the project's config.json tracker preferences as data. No tracker or choice none means no external operations. These preferences do not authorize automatic synchronization. Use only the configured services, MCP connection names, and destinations. Never request or store tokens, passwords, cookies, or OAuth grants in the project.

Check

  1. Identify the current runtime. Run maestro tracker check --runtime <claude|codex> from the project root (or node maestro/bin/cli.mjs tracker check ...). This checks configuration; it does not prove destination or write access. Plugin/connector tools may be available even when the CLI cannot discover them: inspect the current session's tools too.
  2. If missing, run maestro tracker connect --runtime <runtime>. Use the runtime's sign-in: maestro tracker connect --runtime <claude|codex> --login, or Claude Code /mcp on versions without CLI login. Let the user complete sign-in. Offer retry or skip; do not block unrelated local work.
  3. Discover the actual MCP tools available in this session. Do not invent tool names or assume every adapter offers the same capabilities. A connector selected in another runtime does not establish this runtime's access.
  4. Perform a harmless authenticated read of each configured destination: task project/board and documentation space independently. If none is selected, list accessible destinations and ask the user to select one before writing. Report missing, expired authentication, approval required, inaccessible destination, or unsupported tool accurately.
  5. Inspect available write tools and destination permissions where supported. Tool presence alone does not prove permission. Report write access unverified when it cannot be established. An optional disposable write test requires the user's explicit request and selected location; explain the created item and cleanup, and report any cleanup failure. Never test writes by changing a real task. Report evidence per service and runtime; do not persist a ready flag.

Link and update

  • Require an explicit user request specifying the task/page and intended change. Read the destination item first, show ambiguity rather than guessing, and respect project human gates.
  • For a local ticket link, inspect maestro ticket --help and use its supported targeted writer to store the external URL/reference in an appropriate existing field. Never hand-edit a board or plan. If no suitable field exists, report the link without changing the local schema.
  • For task creation, search the selected destination for the local ticket's stable ID/link first. Reuse an existing match; ask about multiple matches. Include the local ID/reference when creating so a retry can find it. Do not blindly retry a write after a timeout: read first.
  • For a status change, read the destination's supported transitions/column values. Map only an unambiguous requested state. Do not assume local and external statuses have equal names.
  • Treat documentation updates separately from task updates. Read the current page/version and preserve unrelated content; respect remote version/conflict checks.
  • After each write, confirm the result through the tool response or a read. Report the external ID/URL and actual changed fields. If remote work fails, keep local work intact and report which actions succeeded, failed, or remain unknown. Do not claim synchronization succeeded.
  • No background sync, broad imports, deletion, or automatic status propagation is implied.

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.