0xmariowu/AgentLint

Check how well your repo supports AI coding agents. Runs 51 evidence-backed core checks across 6 dimensions (Findability, Instructions, Workability, Continuity, Safety, Harness). Seven additional Deep/Session analyzers are opt-in and not run in CI.

View on GitHub

Trust Signals

Scorecard Score
not yet scored
Maintenance Recency
Activelast commit May 28, 2026
License
MIT

Pinned Snippet

workflow.ymlSHA-pinned
uses: 0xmariowu/agentlint@6cf433c4092ee10a3162c2f4ddd68ff13b5d4aff # v1.1.13

tags can be moved; commit SHAs can't. why a SHA?

namedescriptionrequireddefault
project-dirDirectory to scan (default: repo root)no.
fail-belowFail the build if total score is below this threshold (0-100). Default 0 = never fail.no0
formatReport format: terminal, md, jsonl, html, allnoterminal
output-dirDirectory to write reports. Only used when format != terminal.no""
sarif-uploadUpload SARIF to GitHub Code Scanning. Requires the calling workflow to declare `permissions: security-events: write` and (for private repos) a GHAS license. Default: false.nofalse
namedescription
scoreTotal AgentLint score (0-100). Averaged only over dimensions whose checks actually ran — never polluted by a zero for Deep or Session when they are not available in CI.
score-scopeWhich dimensions the total covers: "core" (6 deterministic dimensions, the CI default) or "core+extended" (also includes Deep/Session when opted in).
findabilityFindability dimension score (0-10)
instructionsInstructions dimension score (0-10)
workabilityWorkability dimension score (0-10)
continuityContinuity dimension score (0-10)
safetySafety dimension score (0-10)
harnessHarness dimension score (0-10)