fix(review): severity follows the blocking floor, not the checklist pass or category (GSTA-160)

The checklist bound severity to its passes (Pass 1 CRITICAL, Pass 2 INFORMATIONAL)
and the core floor had no security clause, so GSTA-11's replay saw real bugs filed
INFORMATIONAL. One floor sentence now reaches every reviewer prompt; the passes set
check order only; shell injection covers shell sinks; a state-swap-before-failure
check joins Pass 1 (GSTA-159); the Testing specialist reports a floor-level product
defect as category correctness so classify keeps it blocking.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
This commit is contained in:
garrytanandcapy-ai[bot] committed 2026-10-31 01:11:38 +00:00
1 parent 21277291c8
commit da65a356cf
23 files changed
+211 -82

No files matched your search

+3 -2
View File
@@ -3,6 +3,8 @@
* step whose fixes do not earn a full pass. It reads the fixes, never the whole diff. The parent
* writes every input in literally: a reviewer never sees the parent's shell.
*/
import { BLOCKING_FLOOR } from './review-floor';
export interface FixCheckInput { fixList: string; touchedHunks: string; callers: string; testResults: string }
export const FIX_CHECK_INSTRUCTIONS = `You are the fix-check reviewer for a pre-landing review. An earlier review pass found
@@ -14,8 +16,7 @@ fails open (accepts what it refused) or fails closed (refuses what it accepted),
behaviour the code had before the fix, an error or unreadable state now read as success, a test
made weaker. Name the input or call site that reaches each defect.
Rate each finding against the blocking floor: CRITICAL only when it would fail a test, break
production, or lose data or secrets; everything else is INFORMATIONAL. Output one JSON object per
Rate each finding against the blocking floor. ${BLOCKING_FLOOR} Output one JSON object per
line, {"severity":"CRITICAL|INFORMATIONAL","path":"<file>","line":<n>,"summary":"<what breaks, for which input>"},
or the two words NO FINDINGS when the fixes are sound.`;
+4 -4
View File
@@ -13,6 +13,7 @@ import type { TemplateContext } from './types';
import { CC_BACKGROUND_DEFAULT_SINCE, FOREGROUND_IF_AVAILABLE, BACKGROUND_RECOVERY } from './constants';
import { learningsCapture, LEARNINGS_VERDICT } from './learnings';
import { DEFENSIVE_REVIEW_FRAMING, FIXTURE_SUMMARY_MODE } from './defensive-review';
import { BLOCKING_FLOOR } from './review-floor';
function generateSpecialistSelection(ctx: TemplateContext): string {
const isShip = ctx.skillName === 'ship';
@@ -146,14 +147,13 @@ Every other specialist, including Testing, reads the full diff with fixtures.
Required fields: severity, confidence, path, category, summary, specialist.
Optional: line, fix, fingerprint, evidence, test_stub, advisory, evidence_paths, helper_target, cause.
CRITICAL means a test that would fail, a production break, or a security or data-loss defect; everything else is INFORMATIONAL.
${BLOCKING_FLOOR}
A defect the diff causes but that shows elsewhere (a caller broken by a changed signature, a consumer of a removed export) keeps the victim as \`path\` and names the changed line in \`cause\`: {\\"cause\\":{\\"path\\":\\"file\\",\\"line\\":N}}. Without it the finding reads as pre-existing code.
Leave extraction advice to the core shared-code check; report real defects in duplicated code independently, keep advisory metadata on structural advice, and never call a demonstrated defect advisory because a shared helper could fix it.
If you can write a test that would catch this issue, include it in the \`test_stub\` field.
Use the detected test framework ({TEST_FW}). Write a minimal skeleton — describe/it/test
blocks with clear intent. Skip test_stub for architectural or design-only findings.
For a finding a test could catch, add a minimal \`test_stub\` skeleton in the detected test
framework ({TEST_FW}); skip it for architectural or design-only findings.
Output nothing else: no preamble, no summary or commentary.
+6
View File
@@ -0,0 +1,6 @@
/**
* The blocking floor every gstack reviewer prompt carries (docs/review-policy.md#the-blocking-floor).
* One sentence, so the core checklist, each specialist, the Red Team and the fix-check rate
* severity the same way. test/review-severity-floor.test.ts keeps the markdown copies in step.
*/
export const BLOCKING_FLOOR = 'CRITICAL is the blocking floor: a finding that would fail a test, break production, lose data or secrets, or open a security hole (injection, auth bypass, secret exposure). Everything else is INFORMATIONAL.';