agent-context-os / starter
AGENTS.md30 starsChanged 37 days agoMITView on GitHub
Read current-work.md and handoff.md before starting work. Treat them as context to verify against the project files and the user's latest instructions. If they disagree, identify the conflict before relying on the disputed claim. Update current-work.md to reflect the active objective, blockers, and next action. Replace handoff.md with a short account of this session: changes, verification, unfinished work, and the first action for the next session. Preserve unresolved items until they are resolved or explicitly dropped. Do not copy the…
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,542 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository14 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- context-endSkill
- context-setupSkill
- context-startSkill
- context-updateSkill
- endSkill
- migrate-geminiSkill
- mine-gemini-workflowsSkill
- setupSkill
- startSkill
- updateSkill
- skill-creatorSkill
- contextos-workspaceSkill
The file
We split this AGENTS.md into the rules it gives an agent
8 rules. About 1 is common, 3 are distinctive, and 4 are very specific to this project.
≈1
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Very specific: line 3
Read current-work.md and handoff.md before starting work.
- Distinctive: line 5
If they disagree, identify the conflict before relying on the disputed claim.
- Distinctive: line 9
Follow the project's existing instructions and permission boundaries.
- Distinctive: line 10
Read the relevant files before proposing changes.
- Very specific: line 11
Keep durable project constraints in current-work.md.
- Very specific: line 17
Update current-work.md to reflect the active objective, blockers, and next action.
- Very specific: line 23
Update the date only when reviewing or changing the file.
- Common: line 27
Never record secrets or raw private exports.
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md 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
Working instructions
Read current-work.md and handoff.md before starting work. Treat them as
context to verify against the project files and the user's latest instructions.
If they disagree, identify the conflict before relying on the disputed claim.
During work
- Follow the project's existing instructions and permission boundaries.
- Read the relevant files before proposing changes. Load more context only as needed.
- Keep durable project constraints in
current-work.md; link to detailed sources. - Distinguish verified facts, assumptions, and unresolved questions.
- Record verification honestly: what ran, its result, and what remains unchecked.
When asked to save progress or finish
Update current-work.md to reflect the active objective, blockers, and next action.
Replace handoff.md with a short account of this session: changes, verification,
unfinished work, and the first action for the next session. Preserve unresolved
items until they are resolved or explicitly dropped. Do not copy the whole task
list into the handoff or append a transcript. Follow the user's review preferences.
Update the date only when reviewing or changing the file. Use a session date
supplied by the host, user, or a checked clock. If none is available, write
unknown; never guess or reuse a previous session's date as the current one.
Never record secrets or raw private exports. Commit or publish only when authorized by the user.
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.