agentleFS
alekspetrovalekspetrov/navigator/skills/nav-route/SKILL.md

nav-route

Skill356 starsChanged 13 months agoMITView on GitHub

Print a route report in plain words from the active roadmap or task document and live GitHub state - the goal, the roadmap in phases with ticket numbers, side tracks, the queue now, and what waits for the user. Auto-invoke when user says "route report", "show the route", "where are we on the plan", "status report", "what is left", "what waits for me", or when route mode asks for a report (session start and end, after a dispatch, a merge, a review verdict, or a deviation from the plan).

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

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

    Do NOT invoke for the list of all roadmap documents.

    line 19
  2. Distinctive:

    If the work left the plan since the last report, add one sentence below the report: what left the plan, and why.

    line 96

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Navigator Route

Print a status report that the user can follow without the task document. The report names work by what it does and by its ticket number. It never shows an internal step code.

When to Invoke

  • The user asks for a status: "route report", "show the route", "where are we", "what is left".
  • Route mode is on (route mode on) and one of its events occurs: session start, session end, a dispatch, a merge, a review verdict, a deviation from the plan.

Do NOT invoke for the list of all roadmap documents. That is nav-roadmaps.

The Source: the ## Route Section

The script reads one section of a document. Two kinds of document hold it:

  • A roadmap, .agent/roadmaps/<name>.md: a title, **Goal**: with two or three plain sentences, and the route. A roadmap is the larger unit. Its steps point to task documents and tickets. Write it by hand; it is not a task document.
  • A task document, .agent/tasks/TASK-XX-*.md: the small route of one task. nav-task writes the section for a new task.
## Route

### Phase 1: The gate reads the status line
- [x] 1. Read the status from one shared function. (#48)
- [ ] 2. Show done documents as one count line. (PR #49, needs 1)
- [ ] 3. Install the plugin on the build machine. (operator, needs 2)

### Side tracks
- [ ] Repair the test target on macOS. (#52)

### Waits for you
- [ ] Decide the release number.
  • A step is one checkbox line: an order number and a plain line that says what the step does.
  • The trailing parenthesis is optional. It holds the ticket (#12, PR #12, owner/repo#12), a task document (TASK-12), the word operator for a step that a person does, and needs with the step numbers that must be done first.
  • A step that points to a task document takes its state from the Status line of that document: in progress shows as "In work", done shows as "Done".
  • A step with no ticket and no operator shows as "Not filed".
  • The section holds no state words. The state comes from GitHub when gh answers. A checked box is always done. With no gh, the checkboxes give the state.
  • ### Side tracks holds work outside the main route. ### Waits for you holds the open decisions for the user.
  • The goal comes from **Goal**: (in ## Context of a task document).

The /nav pane draws its route card from the same script (--format json) and lists the route documents behind t. A press there selects the active route; this script follows the selection.

Execution Steps

Step 1: Run the Script

PLUGIN_DIR="${CLAUDE_PLUGIN_ROOT:-$(cat "${NAVIGATOR_CONFIG_HOME:-${XDG_CONFIG_HOME:-$HOME/.config}/navigator}/plugin-root" 2>/dev/null)}"
[ -d "$PLUGIN_DIR/skills" ] || PLUGIN_DIR="$HOME/.claude/plugins/marketplaces/navigator-marketplace"

python3 "$PLUGIN_DIR/skills/nav-route/functions/route_render.py" [--doc .agent/roadmaps/name.md]

With no --doc, the script takes the route selected in the /nav pane, else the first open roadmap, else the last open task document that has a route section. --list prints every route document with its done count. When the user names a roadmap or a task, pass its path.

Choose the form:

MomentCommand
Session start, a plan change, or the user asksthe full form (no flag)
Each other event, when a marker holds an earlier state--short --since <marker file>
The full form with changes marked--since <marker file>
A marker is written (nav-marker)--state (JSON for the ## Route State block)

--since takes a context marker that holds a ## Route State block, or a JSON file from --state. Other flags: --no-gh (no GitHub calls), --format json (the report or the list as data), --queue-label NAME (an issue label that means "in the queue"; default pilot, in-progress, queued; an assignee also counts), --today YYYY-MM-DD.

Step 2: Present

  1. Paste the output as printed. Do not retype the lines and do not add step codes.
  2. If the work left the plan since the last report, add one sentence below the report: what left the plan, and why.
  3. Add nothing else.

Step 3: On an Error

The script exits with code 2 and one line on stderr. It never guesses.

ErrorAction
step N has no plain lineWrite what step N does in plain words in the task document.
step N starts with an internal code 'U2'Replace the code with plain words.
the task document has no '## Route' sectionAdd the section (see examples/route-section.md).
no '**Goal**:' textWrite two or three plain sentences after **Goal**: in ## Context.
needs step N, which the route does not haveCorrect the needs number.

Edit the task document, then run the script again. Ask the user when the plain words for a step are not clear from the document.

Output Format

See examples/route-report.md for the full form and the short form.

Rules

  • The states are: Done (with the date from GitHub), Merged, In the queue now, In work, Next, Later, Blocked: needs step N, Closed, not merged, Closed, not planned.
  • Next is the first open step whose needed steps are done. A step with no needs does not wait for the steps before it.
  • An open pull request, an assigned issue, and an issue with a queue label are in the queue.
  • "What waits for you" lists the operator step that is next, each open pull request, and the open lines of ### Waits for you.
  • The script reads the task document and calls gh api for each ticket. It writes no files.
  • A route covers one document. Other work is one side-track line with its ticket.

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.