skill packProFree PackEditorial Curated5 prompts sequenced

Trust and Verification Workflows

This collection gives editors and technical teams repeatable ways to validate evidence before publishing or acting. The prompts emphasize dated primary sources, untrusted-content boundaries, visual provenance, release-note diffs, and authorized security reproduction with explicit stop conditions.

Use case: Research validation, editorial review, and safe technical analysis
Audience: Editors, analysts, security teams, and AI operators
Tools: ChatGPT · Claude · OpenAI Codex

Production Pipeline

Suggested Execution Order

Agent Skill • TS/JSON
Prompt 01
skill

Source Date Verification Gate

Verify that every time-sensitive claim has a dated primary source, a matching event date, and an explicit freshness assessment.

Agent Skill • TS/JSON
Prompt 02
skill

Prompt Injection Evidence Triage

Separate trustworthy evidence from instructions hidden inside webpages, files, tool output, and other untrusted content.

Agent Skill • TS/JSON
Prompt 03
skill

Visual Claim Provenance Checker

Audit whether an image, chart, screenshot, or generated visual supports the claim attached to it and disclose uncertainty clearly.

Agent Skill • TS/JSON
Prompt 04
skill

Release Note Change Detector

Compare official release notes across dates to isolate product changes, deprecations, migrations, and claims that need retesting.

Agent Skill • TS/JSON
Prompt 05
skill

Authorized Security Reproduction Plan

Translate a public security finding into a bounded, non-destructive reproduction plan with evidence, stop conditions, and remediation checks.

Keep Exploring

More Curated Packs

View all collections