General-purpose coding policy for Baruch's AI agents
91
93%
Does it follow best practices?
Impact
91%
1.15xAverage score across 12 eval scenarios
Advisory
Suggest reviewing before use
tile.json manifest with name, version, summary, and entrypoint (→ README.md). Don't add a separate docs field — keep all documentation in the entrypoint to avoid duplicate tables that drift out of syncREADME.md is the project's README.md — they are the same file. Do not create a separate README for the tile. If the project already has a README, extend it with tile content (rules table, skills table, installation instructions)[](https://tessl.io/registry/<workspace>/<tile>)skills/<name>/SKILL.md, rules live in rules/<name>.mdrules/, skills/<name>/, evals/.tileignore to exclude build artifacts and CI files from the published tiletessl tile lint before every publishCHANGELOG.md will show as orphaned in lint — this is expected; lint only tracks manifest-declared pathssee rules/foo.md), don't duplicate content across rules — if you want to state the same point in two rules, one states it and the other references italwaysApply: true (no other fields needed for rules)# Commit Conventions for commit-conventions.md)tessl skill review --threshold 85 before publishtessl skill review --threshold 85 skills/<name> step to the CI workflow — the review gate is only real if CI enforces itrules/plugin-evals.mdeval-authoring skill — invoke it to generate and curate scenariosWhen you add, remove, or rename a rule or skill, update all of these:
tile.json — add/remove the steering or skill entrytessl skill review step for each skillREADME.md — update the rules table and/or skills tableCHANGELOG.md — add an entry describing the changeAfter modifying rules, audit for cross-rule alignment:
tile.json entries exactlyAfter editing a rule, audit the repo itself against the new rule text and fix any drift in the same PR:
.env.example files, SKILL.md step headings, secret names, etc.) and update them to satisfy the new wording