terraproof
AGENTS.md0 starsChanged 23 days agoMITView on GitHub
These guidelines apply to AI coding agents modifying or extending the Terraproof repository.
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 repository24 other files it gives its agents.
- accuracy-assessmentSkill
- brainstormingSkill
- crs-and-unitsSkill
- design-postgis-tablesSkill
- gdalSkill
- geomasterSkill
- geopandasSkill
- geoparquet-validationSkill
- geospatial-data-discoverySkill
- geospatial-frontendSkill
- geozarrSkill
- networkxSkill
- raster-processingSkill
- sar-basicsSkill
- satellite-preprocessingSkill
- scientific-visualizationSkill
- spectral-indicesSkill
- stac-and-cloud-accessSkill
- statistical-analysisSkill
- systematic-debuggingSkill
- time-series-and-changeSkill
- vector-qualitySkill
- verificationSkill
- writing-plansSkill
The file
We split this AGENTS.md into the rules it gives an agent
12 rules. About 0 are common, 8 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
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
- Distinctive: line 9
Smallest Correct Change: Always prefer the most focused, minimal change that achieves the goal.
- Distinctive: line 11
No Unnecessary Abstractions: Do not introduce abstractions, wrappers, or speculative patterns without a concrete, current requirement.
- Distinctive: line 12
Preserve Working Components: Do not rewrite or refactor working code or content unnecessarily.
- Very specific: line 14
Avoid Rule Duplication: Do not duplicate universal rules across individual skills.
- Distinctive: line 15
Preserve Repository Structure: Do not silently reorganize, add unplanned top-level directories, or alter the repository structure without …
- Distinctive: line 16
Verify Before Completion: Verify all changes, file paths, and syntax before declaring work complete.
- Distinctive: line 17
No False Claims: Never claim tests or verification passed unless they were genuinely executed and verified.
- Very specific: line 18
Preserve Attribution & Licensing: Always maintain upstream licensing information, references, and attribution for any adapted material in …
- Very specific: line 24
Test-Backed Routing Rules: Never add a routing rule or trigger without a corresponding test scenario in tests/router/scenarios.json.
- Very specific: line 25
Register Only Real Skills: Never register nonexistent skills or placeholders as production-ready in routing/registry.json.
- Distinctive: line 26
Precise Triggers Over Broad Keywords: Prefer specific technology and capability triggers over generic terms that cause over-selection.
- Distinctive: line 27
Avoid Domain Over-Selection: Do not route every domain task to every domain skill.
≈ “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
Terraproof — Agent Guidelines
These guidelines apply to AI coding agents modifying or extending the Terraproof repository.
Core Rules
- Smallest Correct Change: Always prefer the most focused, minimal change that achieves the goal.
- Inspect First: Inspect existing work and understand the context before modifying or adding files.
- No Unnecessary Abstractions: Do not introduce abstractions, wrappers, or speculative patterns without a concrete, current requirement.
- Preserve Working Components: Do not rewrite or refactor working code or content unnecessarily.
- Keep Skills Focused: Each skill must address a single domain or task cleanly.
- Avoid Rule Duplication: Do not duplicate universal rules across individual skills; place universal principles in
instructions/. - Preserve Repository Structure: Do not silently reorganize, add unplanned top-level directories, or alter the repository structure without explicit instruction.
- Verify Before Completion: Verify all changes, file paths, and syntax before declaring work complete.
- No False Claims: Never claim tests or verification passed unless they were genuinely executed and verified.
- Preserve Attribution & Licensing: Always maintain upstream licensing information, references, and attribution for any adapted material in
sources/SOURCES.md.
Smart Router Guidelines
- Test-Backed Routing Rules: Never add a routing rule or trigger without a corresponding test scenario in
tests/router/scenarios.json. - Register Only Real Skills: Never register nonexistent skills or placeholders as production-ready in
routing/registry.json. - Precise Triggers Over Broad Keywords: Prefer specific technology and capability triggers over generic terms that cause over-selection.
- Avoid Domain Over-Selection: Do not route every domain task to every domain skill; select only the minimal, highest-confidence tools.
- Concrete Reasons for Dependencies: Hard dependencies (
requires) and recommendations (recommended) must have explicit, documented rationales. - Deterministic Execution: The router must produce the identical skill composition and ordering for identical inputs across all runs.
- Regression Testing for Routing: Any modification to scoring, classification, or registry metadata must pass the entire
npm run test:routersuite with 100% scenario accuracy.
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.