Known issue severity prose consistent

The check is default-on, warning, verify stage. It walks every active KnownIssue (open / reviewed, plus a blank status). A KI whose status is fixed or withdrawn leaves the active set, so it is never flagged. Load failure is a fail. Zero KnownIssues with a parsed prose grade is a pass that inspected nothing of this kind.

Every other severity surface reads only the structured field. Completeness asks that the field exist. High-severity hops scope to high and critical. A quiet downgrade evicts the record from every high-severity check while the human-facing title still says HIGH. No other check reads the title for a self-assigned grade, so the contradiction sits unaudited.

id: KI-cred-inject
title: HIGH credential-injection finding in proxy auth
status: open
severity: low

The parser is precision-first. A severity word counts as the finding's own grade only in three forms: an ALL-CAPS standalone token (HIGH, CRITICAL), a leading grade prefix (Critical:, [High]), or adjacency to the word severity (high severity, severity: high). Incidental hyphenated mentions like high-privilege or low-entropy are ignored. The description is read only in its first 200 runes. When the prose names two different grades, the hop stays silent rather than guess. Recall is partial by design.

proof known-issue edit KI-cred-inject --set-severity high
proof audit --check known_issue_severity_prose_consistent

Either exit clears the hop: make the prose and the field agree, or stamp a severity_bound with a non-empty reviewer and a non-empty rationale. An anonymous note is not a waiver. Filing the KI as fixed to silence a still-open claim is not a close. Deleting the YAML also silences the hop. It destroys the finding rather than reconciling the grade.

The teaching graph is still one active KI whose title says HIGH and whose field says low. Read the finding. Then raise the field, reword the title, or stamp a waiver with a name.

proof known-issue edit KI-cred-inject --set-severity high

The other honest exit is a reviewer-owned bound. Both fields must be non-empty:

severity_bound:
  reviewer: bob
  rationale: prose is legacy triage wording; reproducer bounds real impact to low

Completeness stays on known issue complete. Review currency stays on known issues reviewed. Containment on a security-named KI stays on known issue security remediation present. Do not treat a Salesforce bulletin as this cell. Jama still authors. Proof vs Jama.

Warning severity. A counted finding does not block the audit. The hop does not parse a CVSS vector. It does not fetch the tracker, re-run a PoC, or confirm the security claim is true. It does not treat high-privilege or low-entropy as a grade. Two different grade words in the same prose stay silent. A grade buried past the first 200 runes of the description is not this finding's own grade. A KI with no parsed prose grade is skipped. Zero KnownIssues with a prose grade is a pass that inspected nothing of this kind. Load failure is a fail. An unsigned severity_bound is not stamped. A warning is not a waiver: proof waive is a human authorization gate. The hop does not prove the Go. We have not scored this floor against a frozen Jama pack, a GitHub advisory export, or a Salesforce status page. The loss is named, not scored.

The completeness hop stays on known issue complete. The review-date hop stays on known issues reviewed. The containment hop stays on known issue security remediation present. The engagement stays on software correctness audit. Jama still authors.

What is known issue severity prose consistent? Same question. Same URL.

Is this known issue complete? No. That hop is whether the record has evidence and an origin. This hop is whether the prose grade and the field name the same tier. See known issue complete.

Is this known issues reviewed? No. That hop is whether review_date is still current. This hop is whether the title and the field contradict. See known issues reviewed.

Is this known issue security remediation present? No. That hop is whether a security-relevant KI documents containment. This hop is the prose-vs-field grade. See known issue security remediation present.

Is this high severity reproducer grade? No. That hop is whether a high or critical KI carries a runnable witness. A quiet downgrade can hide the KI from that hop. This hop names the downgrade. See high severity reproducer grade.

Is this a Salesforce known issue? No. Ads known issue is that status page. A hosted vendor bulletin is not this cell.

Does high-privilege fire? No. Incidental hyphenated mentions are ignored. The hop needs an ALL-CAPS token, a leading grade prefix, or adjacency to severity.

Does a title that says HIGH and LOW fire? No. Two different grade words are ambiguous. The hop stays silent rather than guess.

Does a missing known-issues directory fail? No. Zero KnownIssues with a prose grade is a silent pass. Put a YAML there when you disclose, then keep the title and the field on the same tier.

Does a green hop prove the Go? No. The hop does not run the PoC. It does not fetch the tracker. It does not prove the function.

Is Proof a Jama alternative for the shall? No. Jama still authors. Proof vs Jama.