메뉴
Triggers a secondary verification pass for any agent output containing factual claims, numbers, dates, or named entities before the output is acted on
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | fact-check-before-trust |
| version | 1.0 |
| category | core |
| description | Triggers a secondary verification pass for any agent output containing factual claims, numbers, dates, or named entities before the output is acted on |
verification-before-completion checks that a task was done. This skill checks that the facts are correct. An agent confidently reporting a £716 visa fee as £70,000 will pass completion verification — this skill catches it.
Invoke this skill before treating any agent output as authoritative when the output contains:
Skip for: code output, file system operations, and clearly self-contained tasks (renaming a variable, formatting a document).
Step 1 — Extract claims Identify every verifiable claim in the output. List them explicitly:
Claim 1: UK visa fee is £716
Claim 2: Processing time is 3 weeks
Claim 3: Applies to Tier 2 (Skilled Worker) visa category
Step 2 — Score each claim For each claim, assign a confidence level:
Step 3 — Verify low/medium claims For each Low or Medium claim:
Step 4 — Classify output
| Result | Meaning |
|---|---|
| ✓ Verified | All claims High confidence |
| ⚠ Uncertain | 1+ Unverifiable claims |
| ✗ Contradicted | 1+ claims conflict with found evidence |
Step 5 — Surface to user
verification-before-completion checks: "Did the agent do the task?" (task completion)
fact-check-before-trust checks: "Is what the agent said true?" (output accuracy)
Both should be used for research, financial, legal, and compliance workflows.
Handles PR review feedback by fetching comments, grouping issues, fixing one group at a time, and verifying before replies.
Reviews whether a skill will trigger reliably, guide useful behavior, avoid overlap, and produce testable outcomes.
Diagnoses OpenClaw provider, fallback, channel, MCP, and gateway config issues with read-only scans and stateful summaries.
Tracks required validation gates, records pass/fail/waived results, and reports readiness before task completion.
YAML-based delegation grant ledger — issues, validates, and tracks scoped permission grants for sub-agent expansions with token budgets and auto-expiry.
Walks the memory DAG to recall detailed context on demand — query, expand, and assemble cited answers from hierarchical summaries without re-reading raw transcripts.