recoup-content-reactive-post
Pass
Audited by Gen Agent Trust Hub on Jul 17, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The agent utilizes standard shell commands (curl, jq, sed) to interact with platform services and organize artist-specific files within the project directory.
- [EXTERNAL_DOWNLOADS]: The skill connects to authorized endpoints (api.recoupable.dev) for artist research, milestone tracking, and media generation. It also documents the use of yt-dlp as a tool for retrieving audio files.
- [PROMPT_INJECTION]: The skill implements automated quality checks via a Stop hook and the analyze-gate process. These instructions serve as functional constraints that require the agent to verify visual outputs through a dedicated analysis tool, preventing false claims about asset readiness. The static analysis hint regarding concealment refers to this legitimate verification step.
- [PROMPT_INJECTION]: The skill processes data from external sources, such as web trend summaries and artist career feeds, to inform content creation. 1. Ingestion points: Career event feeds and web search summaries retrieved via the research API in references/research-context.md. 2. Boundary markers: Explicit delimiters for separating external research data from agent instructions are not defined. 3. Capability inventory: The skill possesses capabilities for shell command execution and local file system modifications across all referenced scripts. 4. Sanitization: The agent relies on automated parsing of structured data, though the final narrative output is generated based on the raw research text.
Audit Metadata