adaptive-engineering
AGENTS.md1 starsChanged 11 days agoMITView on GitHub
Use the least process necessary to produce the simplest correct, verified solution. For every meaningful software task: 1. Classify it as FAST, STANDARD, or DEEP based on risk, scope, ambiguity, reversibility, and blast radius. 2. Inspect existing code and native capabilities before adding dependencies or abstractions. 3. If .adaptive/memory/claims.json exists, treat it as re-verifiable context only: repository evidence is the source of truth, changed evidence makes claims stale, and stale claims must be reverified before reuse. 4. Use a complexity…
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 31,774 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 repository10 other files it gives its agents.
- adaptive-engineeringSkill
- code-reviewSkill
- existing-system-scanSkill
- memory-groundingSkill
- minimal-planningSkill
- simplification-gateSkill
- systematic-debuggingSkill
- task-triageSkill
- verificationSkill
- verified-implementationSkill
The file
We split this AGENTS.md into the rules it gives an agent
8 rules. About 0 are common, 7 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈7
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
- Distinctive: line 3
Use the least process necessary to produce the simplest correct, verified solution.
- Very specific: line 8
If .adaptive/memory/claims.json exists, treat it as re-verifiable context only: repository evidence is the source of truth, changed …
- Distinctive: line 9
Use a complexity budget as a tripwire, not a target.
- Distinctive: line 11
Verify with executable evidence.
- Distinctive: line 12
Run a simplification pass before completion.
- Distinctive: line 13
When memory exists or durable knowledge changed, run the memory consistency gate and refresh only affected claims.
- Distinctive: line 21
Never add a database, vector store, daemon, MCP server, or external service merely to support Adaptive Engineering memory.
- Distinctive: line 23
Never lower safety, security, data-integrity, compatibility, or production-critical checks just to fit a requested mode.
≈ “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
Adaptive Engineering — portable instructions
Use the least process necessary to produce the simplest correct, verified solution.
For every meaningful software task:
- Classify it as FAST, STANDARD, or DEEP based on risk, scope, ambiguity, reversibility, and blast radius.
- Inspect existing code and native capabilities before adding dependencies or abstractions.
- If
.adaptive/memory/claims.jsonexists, treat it as re-verifiable context only: repository evidence is the source of truth, changed evidence makes claims stale, and stale claims must be reverified before reuse. - Use a complexity budget as a tripwire, not a target.
- Implement the smallest coherent solution.
- Verify with executable evidence.
- Run a simplification pass before completion.
- When memory exists or durable knowledge changed, run the memory consistency gate and refresh only affected claims.
FAST: local, low-risk, reversible change. Inspect -> relevant memory check -> reuse -> minimal change -> targeted verify -> simplify.
STANDARD: subsystem-level feature/bug/refactor. Inspect -> reverify relevant claims -> acceptance criteria -> short plan + budget -> implement/test -> simplify -> verify -> memory consistency.
DEEP: security/auth, data migration, public contracts, CI/CD/production infrastructure, cross-service consistency, difficult rollback, or major architecture. Discover invariants + relevant claims -> reverify and search for contradictions -> requirements/failure modes -> trade-offs -> staged plan + rollback + budget -> implement/test -> independent review -> simplify -> end-to-end verify -> drift check.
Never add a database, vector store, daemon, MCP server, or external service merely to support Adaptive Engineering memory. Plain repository-local files must remain sufficient.
Never lower safety, security, data-integrity, compatibility, or production-critical checks just to fit a requested mode.
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.