template-intake
Skill2 starsChanged 12 days agoMITView on GitHub
Take in upstream suggestions recorded by brains made from this template: collect open entries with _meta/brain upstream list, accept or decline each with the person, make accepted changes through brain-write-skill or brain-write-script with tests, and answer every one in _meta/CHANGELOG.md by id. Use when asked in a template session to check brains for suggestions or to process upstream feedback. Does NOT write into any brain, and does NOT commit, which is brain-sync.
In this repository10 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- brain-retroSkill
- brain-setupSkill
- brain-syncSkill
- brain-upstreamSkill
- brain-write-modSkill
- brain-write-practiceSkill
- brain-write-scriptSkill
- brain-write-skillSkill
- template-make-brainSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 4 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
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
- Very specific: line 26
Run this in the template's own checkout: brain.json role is template and local/ exists.
- Distinctive: line 57
Write the change for any brain, with no hostname, project, person or path taken from the suggesting brain: everything here ships, and the …
- Distinctive: line 87
Do not commit here.
- Distinctive: line 91
Write, stage or commit in a brain, or run anything there that writes: the brain's own sessions change the brain, and the template only …
- Very specific: line 98
Every open, unanswered id upstream list showed has exactly one - landed: or - declined: line in _meta/CHANGELOG.md, under a new heading …
- Distinctive: line 104
No file in any brain was written, and no brain's private detail is in the template.
≈ “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.
Files
Take in suggestions from brains
A brain cannot change the files the template owns, so it writes what one of them needs as a
suggestion in its own docs/UPSTREAM.md. This skill reads those, decides each one with the person, makes the
accepted changes here, and answers every one in _meta/CHANGELOG.md; each brain takes the answer in
with its next update. Reading is the only thing this skill does to a brain.
Where this sits
<!-- where:begin — generated by _meta/brain routing sync; do not edit -->Runs after: nothing — an entry point into the tree.
Hand to next: brain-sync · brain-write-script · brain-write-skill
<!-- where:end -->Call it directly when: in a template session, the person asks to check the brains for
suggestions, or brings one from a brain local/brains.txt does not list; nothing runs first.
1. Collect
Run this in the template's own checkout: brain.json role is template and local/ exists.
_meta/brain upstream list reads every brain listed in local/brains.txt. For a brain not listed
there (a clone of one kept on another machine, say), pass its path:
_meta/brain upstream list "<brain>". Each line is <brain> <id> <status>, plus the template's
answer when CHANGELOG.md already has one. Work the open ids that have no answer. An id the
changelog answers is done, even while the brain still reads open: it flips when that brain
updates. A listed brain that is no longer there: tell the person; local/brains.txt is theirs to
tidy.
For each id, read its entry in that brain's docs/UPSTREAM.md: - template: (the version it was
written against), - files:, ### What happened, ### Proposed change. Read only: never edit,
stage or commit anything in a brain, not even the status line, which the brain flips itself in the
commit that brings the answer in.
2. Decide each one with the person
One at a time: say what the brain hit, what it proposes, and what the template does now (the files may have changed since the version the entry names, and it may already be fixed). Recommend accept or decline, with the reason, and let the person decide. Accept a change that helps any brain; decline one that fits only that brain's work, undoes a rule the template keeps on purpose, or is already fixed (say by what). A partial accept is an accept whose summary says what landed. Two brains with the same problem get one change and an answer each.
3. Make each accepted change
- A
brain-*skill,_meta/ROUTING-META.mdor_meta/gates.json: throughbrain-write-skill. - The harness (
_meta/brain,_meta/brainlib/,_meta/managed/,_meta/skeleton/): throughbrain-write-script, as its "In the template" section says.
Hand the proposal on as the starting point, not as the change: it was written from one brain's work. Write the change for any brain, with no hostname, project, person or path taken from the suggesting brain: everything here ships, and the template is public. Run the suite before answering:
PYTHONDONTWRITEBYTECODE=1 python3 -m unittest discover -s tests -t .4. Answer every one in the changelog
Bump _meta/VERSION once for this intake: the patch number when every accepted change fixes
something (0.1.0 → 0.1.1), the minor number when any adds behaviour (0.1.0 → 0.2.0). Add a
heading for the new version above the previous one in _meta/CHANGELOG.md, and under it one line
per id worked, in exactly this shape, because upstream list and each brain's update parse it:
## <version> — <YYYY-MM-DD>
- landed: <id> — <what changed, in one line>
- declined: <id> — <the reason, in one line>The separators are em dashes (—) and the date is today's. Every id worked gets exactly one line;
one the person wants to leave for later stays open, with none. An intake that only declines still opens a patch heading, so each brain's next update reports the
answer as new. Write the summaries and reasons for any brain too: they reach every brain.
5. Hand to brain-sync
brain-sync commits in the template, running the test suite first: each change under its own scope
(skill(<name>), meta(<area>), test(<area>)), with _meta/VERSION and _meta/CHANGELOG.md.
Committing is enough, with no release step: brains update from the template's committed history.
Do not commit here.
Never
- Write, stage or commit in a brain, or run anything there that writes: the brain's own sessions change the brain, and the template only reads it.
- Answer an id without the person's decision, or work one and leave it without a changelog line.
- Carry a brain's private detail (a host, a project, a person, a local path) into the template.
Done when
- Every open, unanswered id
upstream listshowed has exactly one- landed:or- declined:line in_meta/CHANGELOG.md, under a new heading whose version matches_meta/VERSION, or the person chose to leave it open for later. - Each answer was the person's decision.
- Each accepted change was made through
brain-write-skillorbrain-write-script, and the test suite passes. - No file in any brain was written, and no brain's private detail is in the template.
- The work is handed to
brain-sync; nothing was committed here.
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
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.

