seek-verdict: Blind Verification for P2 Dismiss

When NOT to Use

• General code review (use /codex-review-fast)
• P0/P1 findings (must fix per fix-all-issues.md)
• Nit findings (use [NIT_DEFERRED])
• Architecture debates (use /codex-brainstorm)

Scope

Severity	Allowed	Mechanism
P0/P1	No — must fix	fix-all-issues.md
P2	Yes	This skill
Nit	No — use existing	[NIT_DEFERRED]

3-Phase Protocol

sequenceDiagram
    participant C as Claude
    participant S as seek-verdict
    participant X as Codex (fresh thread)

    C->>S: Finding + dismiss intent
    Note over S: Phase A: Candidate Packaging
    S->>S: Extract finding_packet<br/>(no Claude conclusions)
    S->>X: Phase B: Blind Independent Verdict
    Note over X: Independent research<br/>git status / grep / cat
    X-->>S: verdict + confidence + evidence
    Note over S: Phase C: Policy Mapping
    S-->>C: Result + audit log

Phase A: Candidate Packaging

Extract finding artifact from review output:

Field	Source
finding_key	file + canonical_issue_text
severity	P2
original_finding_text	Codex review original (secrets redacted)
origin_thread_id	Review session threadId
current_head_sha	git rev-parse HEAD
relevant_diff	git diff HEAD -- <file>

Critical: Record Claude's dismiss hypothesis locally. Never send it to Codex.

Phase B: Blind Independent Verdict

Use the prompt template in Verdict Prompt.

Requirement	Detail
Thread	Fresh mcp__codex__codex (never reuse review thread)
Sandbox	read-only
Approval policy	never
Anti-anchoring	No Claude conclusions in prompt

Phase C: Policy Mapping

Apply thresholds from Policy Mapping.

Result	Condition	Next Action
DISMISS_VERIFIED	NON_ACTIONABLE + confidence >= 0.80 + evidence >= 2	Log [DISMISS_VERDICT], continue
FIX_REQUIRED	ACTIONABLE + confidence >= 0.70	Return to fix loop
NEED_HUMAN	UNCERTAIN or low confidence	Stop, escalate

Output [DISMISS_VERDICT] audit trail (format in references/policy-mapping.md).

Rebuttal

If Codex returns FIX_REQUIRED and Claude has objective counter-evidence:

• 1 round max via mcp__codex__codex-reply (same verdict thread)
• Only objective artifacts (tests, specs, language semantics)
• After rebuttal: still FIX_REQUIRED -> fix; ambiguous -> NEED_HUMAN

Anti-Abuse Guard

See Policy Mapping for thresholds and session scope.

• 3 consecutive DISMISS_VERIFIED -> [DISMISS_PATTERN_WARN] + heightened thresholds
• Session end or branch switch resets counter

Output

[DISMISS_VERDICT] key=<file|canonical_issue> | severity=P2 | verdict=<DISMISS_VERIFIED|FIX_REQUIRED|NEED_HUMAN> | confidence=<0..1> | codex_thread=<id> | evidence=<brief> | timestamp=<ISO8601>

Verification

• Finding is P2 severity (P0/P1/Nit rejected)
• Codex prompt contains no Claude conclusions (anti-anchoring)
• Fresh Codex thread used (not review session thread)
• [DISMISS_VERDICT] audit trail output with all required fields
• Anti-abuse streak tracked (warning at 3 consecutive dismissals)

References

File	Purpose
Verdict Prompt	Codex blind verification prompt template
Policy Mapping	Confidence thresholds, audit format, anti-abuse