laohan-chuangzuo
Pass
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external, potentially untrusted data from video transcriptions and web page content, creating a surface for indirect prompt injection attacks. \n
- Ingestion points: The skill ingests data from video files via transcription (Pre-A) and web page content via URL grabbing (Pre-B). \n
- Boundary markers: Implements a structured "Layer A/B" system to separate raw facts from AI-generated analysis, as described in
references/skill.md. It also utilizesfact_boundaryand SHA256 binding in its creation contracts. \n - Capability inventory: The skill possesses significant capabilities including shell execution (
ffmpeg,say,ffprobe,pdftoppm), local file system access (writing tooutput/,episodes/, and the user's desktop), and network access (curl,huashu-chrome). \n - Sanitization: Uses structured data isolation and the "Sepia" language review process (
references/sepia.md) to audit and filter content before finalizing the output. \n- [COMMAND_EXECUTION]: The skill executes multiple system commands to process media and validate its own outputs. \n - Evidence:
SKILL.mdusesffmpegfor audio extraction andpdftoppmfor slide generation.scripts/check-script-contract.mjsusessayfor text-to-speech validation andffprobefor audio duration measurement. \n- [EXTERNAL_DOWNLOADS]: The skill interacts with external APIs and well-known services to fetch and process data. \n - Evidence:
references/transcription.mdusescurlto interact with the SiliconFlow API.references/dual-engine-gemini.mddescribes a workflow using Google Gemini.SKILL.mduses the Jina Reader service to fetch web content. These are recognized services or trusted organizations. \n- [DYNAMIC_EXECUTION]: The skill executes local Node.js scripts to perform complex workflow validations. \n - Evidence: Multiple Node.js scripts in the
scripts/directory, such ascheck-script-contract.mjs, are executed to verify SHA256 hashes and ensure the integrity of the creation process.
Audit Metadata