Crawl-finding disposition · Playbook
Crawl-finding disposition — high-level pointers (local fallback)
bisharp tells you, per finding, fix / dismiss / needs-human, with a stated reason. The detailed,
current disposition standards — which findings are safe to auto-dismiss by URL role, how the
needs-human cap should move as a rule set matures, and how to read a --calibrate disagreement — live
in the team wiki (set OUTLINE + OUTLINE_API_URL and the report links them directly). This file is
the minimal high-level fallback for when the wiki isn’t configured: direction only, deliberately
not step-by-step, so there’s no detailed content to drift out of sync with the wiki.
- noindex — correct and dismissable on non-money URLs (paginated, archive, cart, thank-you); a fix on anything that can earn, because it removes the page from Google.
- canonical-missing — dismissable off money pages; a real fix on a page that can earn, since it is exposed to duplicate-URL splitting.
- h2-duplicate / spelling / grammar — template artefacts and dictionary noise, not defects; dismiss with the reason on record rather than re-litigating them per URL.
- thin-content — never auto-dismiss on a money page; word count alone can’t tell a genuine gap from a page that’s short by design, so it routes to a human (and to bibarel).
- title-length — a proxy for a pixel budget, not the defect itself; route to orbeetle before spending time rewriting to a character count.
- redirect chains / internal 3xx / broken 4xx / server 5xx — always fix; severity order matters (5xx first), and each routes to the skill that actually repairs it (abra, klink).
- needs-human cap — treat a run that blows through the cap as a rule-set gap, not a crawl problem; the unmapped finding names are the backlog of rules to add next.
- Calibration —
--calibratemeasures agreement against a real export; a run under 80% means the rules are not yet safe to trust unattended, not that the crawl is wrong.
Full playbook → the team wiki’s crawl-triage disposition standards.