agentleFS
flurdyflurdy/agent-skills/skills/name-session/SKILL.md

name-session

Skill9 starsChanged 3 months agoMITView on GitHub

Derive a conventional session name from the branch ticket, active bead, open PR, and current work, then print the paste-ready rename command for the active client. Use when a session's auto-name is generic.

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 3 are distinctive, and 6 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    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. Very specific:

    Never use the shell, PATH, filesystem, process list, or installed binaries to detect another client.

    line 19
  2. Very specific:

    Treat the client as unknown when that surface does not conclusively identify Pi or Claude Code.

    line 20
  3. Very specific:

    Pi:: use /name {session-name}. Never suggest /rename or /settings name in Pi.

    line 24
  4. Very specific:

    Claude Code: use /rename {session-name}.

    line 25
  5. Distinctive:

    If the client is unknown, state both commands rather than guessing.

    line 26
  6. Very specific:

    Read the current branch with git rev-parse --abbrev-ref HEAD, using the repository wrapper where required.

    line 47
  7. Very specific:

    If neither applies and a PR is clearly this session's topic, use its known identity or a scoped gh pr view --json number,title.

    line 54
  8. Distinctive:

    Keep the derivation note to one short clause.

    line 81
  9. Distinctive:

    Don't ask the user to confirm the name before printing it — printing *is* the proposal, and they can edit the line before pasting.

    line 86

≈ “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

Name session — propose a conventional session name

Build a <scope>-<descriptive> name from the current context and emit the active client's ready-to-paste rename command.

Important — client command and limitation

Harness selection comes from the current tool surface. Never use the shell, PATH, filesystem, process list, or installed binaries to detect another client. Treat the client as unknown when that surface does not conclusively identify Pi or Claude Code. These rules are shared with /wrap-up and /handoffs; command printing is a proposal, not execution.

  • It cannot rename the session for you. Slash commands emitted in model output are inert text; the user must enter the command in the client's command input.
  • Pi: use /name {session-name}. Never suggest /rename or /settings name in Pi.
  • Claude Code: use /rename {session-name}.
  • If the client is unknown, state both commands rather than guessing.

Convention

<scope>-<descriptive>
  • scope — the most specific identifier available, in priority order:
    1. Jira ticket from the branch name (e.g. AB-1505) — keep its natural Jira case.
    2. Active bead ID (e.g. bd-123) if no ticket.
    3. PR number (e.g. pr-6563) if neither.
    4. Omit the scope entirely if none apply — just use the descriptive part.
  • descriptive — kebab-case, ≤4 words, the most specific noun phrase for what this session is doing. rebase-pr-status beats git-stuff. Derive it from the conversation, not the branch (the branch already gives the scope).

Examples: AB-1505-rebase-pr-status, bd-412-flaky-test-hunt, pr-6563-review-comments, auth0-logout-investigation.

Instructions

1. Gather scope signals

Read the current branch with git rev-parse --abbrev-ref HEAD, using the repository wrapper where required. Extract a ticket with [A-Z]+-\d+; detached HEAD is not a branch/ticket signal.

Without a ticket, prefer the bead explicitly identified as this session's work. If tracker validation is needed, run ~/.agents/skills/next/scripts/next-select resolve <selector> and use the returned owner for bd -C <directory> show <id>. Unavailable/ambiguous ownership drops that signal; do not scan all in-progress claims and assume one belongs here. Existing known session context needs no fetch. If neither applies and a PR is clearly this session's topic, use its known identity or a scoped gh pr view --json number,title. No speculative PR request.

Fail soft: unknown evidence stays unknown. Outside Git, known session context can still name the work; otherwise use the descriptive part alone. No tracker, Git, or client state is changed.

2. Derive the descriptive half

From the current conversation, pick the ≤4-word kebab phrase that best names what this session is for. Prefer the concrete task over the topic — rebase-pr-status over maintenance. If the session genuinely spans several unrelated things, name the dominant one; don't try to cram them all in.

3. Emit the rename command

Compute {session-name} first: {scope}-{descriptive} when scope exists, otherwise just {descriptive} with no leading hyphen. For Pi, render exactly:

**Proposed session name** — scope from {where the scope came from}, descriptive from this session:

/name {session-name}


Paste it into Pi's command input and press Enter.

For Claude Code, substitute /rename for /name. If the client is unknown, provide both commands and label them by client.

Keep the derivation note to one short clause. If you had to fall back (no ticket, no bead), say which fallback you used so the user can override.

Notes

  • Pairs with /wrap-up and /handoffs, which should emit the same client-specific rename command for end-of-session and resume respectively. This skill is the mid-session, on-demand version.
  • Don't ask the user to confirm the name before printing it — printing is the proposal, and they can edit the line before pasting.

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.