agentleFS
adamlinscottadamlinscott/claude-skills/skills/where-are-we/SKILL.md

where-are-we

Skill3 starsChanged 16 days agoMITView on GitHub

[Adam''s Skills] Quick status of the current session: what is done and what is left. Checks how long it has been since the last message before this one. If the user has been active within the last half hour, it answers with just the two lists. If they have been away longer, or the gap cannot be measured, it opens with a short paragraph on what the task actually is. Grounds the lists in git and files on disk, not memory. Report only: it changes nothing and does not resume work. Triggers on /where-are-we, or on the phrase "where are we" anywhere in the user''s message, with or without a question mark.

2 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

7 rules. About 0 are common, 6 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    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. Distinctive:

    Read the real state, not your memory of it: git status, git diff --stat against the base branch, recent git log, the active todo list, and …

    line 33
  2. Distinctive:

    Do not call a piece of work done without confirming it is on disk.

    line 35
  3. Distinctive:

    When away, two more things, only when they have content:

    line 51
  4. Very specific:

    Follow the *Sentence-level shaping* rules in the ttp skill (../ttp/SKILL.md, or ~/.claude/skills/ttp/SKILL.md when installed) for both.

    line 59
  5. Distinctive:

    No closing question, no offer to continue.

    line 64
  6. Distinctive:

    Half-applied edit or incoherent working tree: Put it first under *Left*, plainly.

    line 71
  7. Distinctive:

    Compacted session: Write from what survives, and say in one line where the detail runs out.

    line 72

≈ “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

Where Are We

The user wants to know how far the work in this session has got. How much they need depends on how long they have been gone, so measure that first.

Tune here: the user counts as still active if the last message before this one was under 30 minutes ago.

1. Measure the gap

node <skill-dir>/idle.mjs ${CLAUDE_SESSION_ID}

<skill-dir> is the folder this file is in. Resolve it from this file's real path: ~/.claude/skills/where-are-we is usually a link, and the script sits beside its target. It prints one line of JSON:

  • {"lastMessageAt": ..., "minutesAway": n} — the gap between now and the last message before the user's current one.
  • {"unknown": "<why>"} — no transcript, or no earlier message. Treat as away.

Under the threshold: active. At or over it, or unknown: away. If in doubt, choose away.

2. Check the state before you write

Read the real state, not your memory of it: git status, git diff --stat against the base branch, recent git log, the active todo list, and any plan or spec file written this session. Do not call a piece of work done without confirming it is on disk. If a claim comes only from the conversation and cannot be checked, mark it (recalled).

3. The answer

Active — two short lists, nothing else:

Done — completed work, up to five bullets.

Left — what remains, in the order it will be done, up to five bullets. If something is waiting on the user, say so on that bullet. If nothing is left, say Nothing left.

Away — the same two lists, preceded by one short paragraph (two to four sentences) saying what the task is: what is being built or fixed, and why. Write it for someone who has forgotten the session: no "as we discussed", no session-coined shorthand left unexplained, real file names.

When away, two more things, only when they have content:

  • Changed while you were away — one or two plain sentences if the repo moved under them: new commits on the branch, a base branch that advanced, files they did not change. Goes after the paragraph, before the lists.
  • Update footnote — if /check-skill-updates is installed, run it after the lists and append its one line if it has one. Skip it if Left opens with a half-applied edit.

Follow the Sentence-level shaping rules in the ttp skill (../ttp/SKILL.md, or ~/.claude/skills/ttp/SKILL.md when installed) for both.

4. Stop

No closing question, no offer to continue. If the user's message asked for something else as well, answer that after the status.

Edge cases

  • Nothing has happened yet in this session. Say so in one line, and give the repo state from step 2 only if there is uncommitted work.
  • Half-applied edit or incoherent working tree. Put it first under Left, plainly.
  • Compacted session. Write from what survives, and say in one line where the detail runs out.

Not this skill

/reground also takes stock, but it distrusts the session and re-anchors the agent to code evidence. /where-are-we trusts the session and re-orients the user. If the work itself has gone off the rails, that is /reground.

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.