community-post
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from local
plan.mdfiles to provide context for AI tasks, representing a potential data ingestion surface. - Ingestion points: Reads video title and context from
./youtube/episode/[episode_number]_[topic_short_name]/plan.md. - Boundary markers: The orchestrator does not define explicit delimiters or instructions to ignore embedded commands in these files.
- Capability inventory: The skill can read local files, invoke external specialized skills, and append content to a local
community-posts.mdfile. - Sanitization: No sanitization or filtering of the input file content is performed by the orchestrator.
- [SAFE]: No malicious patterns such as credential theft, obfuscation, or unauthorized network activity were detected. All file operations and external references are scoped to the project's functional requirements and trusted vendor resources.
Audit Metadata