Validate dogfood evidence against golden rule and manifest for production claims.
npx skills add https://github.com/usetheodev/theo-ui --skill dogfood-usetheodev
Please help me install this Agent Skill. Skill: dogfood Source: https://github.com/usetheodev/theo-ui/tree/main/.claude/skills/dogfood Command: npx skills add https://github.com/usetheodev/theo-ui --skill dogfood-usetheodev
Honesty gate that blocks any production-ready / production-grade / v1.0 claim without recorded evidence of sustained internal use. Reads the project's dogfood manifest at knowledge-base/dogfood/manifest.md plus evidence files at knowledge-base/dogfood/evidence/. Applies the project-specific dogfood golden rule (rules/dogfood-golden-rule.md — created per project), and emits EVIDENCE_SUFFICIENT / EVIDENCE_WITH_CAVEATS / EVIDENCE_INSUFFICIENT. Invoke before any production claim, milestone promotion targeting v1.0, README/CHANGELOG edits that touch status language, or release decision.
Run the dogfood audit to validate the current evidence against the golden rule and manifest.
High-intent search queries and answers about installing and using this skill.
Production-readiness governance is enforced by auditing recorded dogfood evidence against a project-specific golden rule and manifest, emitting a formal verdict that blocks premature v1.0 or production-grade claims.
A dogfood evidence manifest is a structured log tracking sustained internal use of a project. You need to maintain one before making any production-ready, production-grade, or v1.0 status claim in READMEs or release decisions.
You audit dogfooding evidence by running the audit mode, which validates evidence files in the knowledge base against the project's golden rule and manifest, returning EVIDENCE_SUFFICIENT, EVIDENCE_WITH_CAVEATS, or EVIDENCE_INSUFFICIENT.
Yes, you need to create a project-specific golden rule file at rules/dogfood-golden-rule.md to define the criteria for release readiness before the audit can validate your evidence.
The best way to log internal usage evidence is using the log-evidence mode, which appends sustained usage records to the evidence directory for subsequent audit validation against the manifest.
A production-ready claim fails the readiness audit when recorded evidence of sustained internal use is insufficient or missing, resulting in an EVIDENCE_INSUFFICIENT verdict that blocks the status claim.
LuminaVaultCompile raw markdown files into structured summaries and concept articles.
dianlightRead completed tasks from docs/tasks/ and append entries to CHANGELOG.md.
h4vzzConfigure CI/CD pipelines with linting, tests, Docker builds, and staged deployments.
tunnelevenRun verification commands and review output before claiming completion.
