Scheduled sweep that checks a repository against its team standards and reports one Tessl finding per standard it drifts from, for the `remediate` action to fix. Use as the skill of the weekly conformance-sweep schedule.
68
85%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
Passed
No findings from the security scan
Report drift; do not fix it. Each finding raises tessl:finding.created, which starts remediate, so the fix goes through the same review path as any other factory pull request.
TESSL_PROJECT_ID, supplied as a schedule input: this repository's Tessl project.rules/*-standards.md) and the AGENTS.md or CLAUDE.md chain. Only written rules count.genproto/, pb/), vendored files and agent context (.claude/, .agents/, .tessl/).tessl finding list --project $TESSL_PROJECT_ID --limit 100 --json, without --state, following links.next with --after. Match on the standard and the files, not on title alone, and view likely matches before deciding. Skip a standard whose matching finding is ignored. For one that is open, the drift is still there: append this run's evidence as a new observation with tessl finding update <id> --file changes.json --json, sending only observations (no loopId or runId), and do not create another.tessl finding create --project $TESSL_PROJECT_ID --file finding.json --json:{
"schemaVersion": 1,
"title": "Conformance: <standard, in a few words>",
"observations": [
{
"observedAt": "<now, ISO 8601>",
"evidence": ["<path>:<line> <what violates the rule>", "..."]
}
],
"analysis": "<the rule, quoted, and how these lines break it>",
"recommendation": "<the specific change that makes them conform>",
"expectedImprovement": "<what conforming gives the team, without inventing a measured benefit>"
}Leave out loopId and runId; the service derives them from the scheduled run. Each evidence string is one observed fact.
caafac3
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.