product0-preparing-brief
Pass
Audited by Gen Agent Trust Hub on Jul 13, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection because it is designed to ingest and process data from external files that may contain instructions. While the skill aims to synthesize these into a developer brief, the agent could be influenced by malicious instructions embedded in the source content.
- Ingestion points: The skill requires the agent to read the "entire canonical brief" and "approved Product0 content" from the project environment (SKILL.md).
- Boundary markers: The instructions specify a "Hard boundary" to prevent implementation planning, but do not provide explicit security boundaries or instructions to ignore embedded commands within the processed data.
- Capability inventory: The skill frontmatter explicitly requires "project file read/write access" to update the brief in place.
- Sanitization: There are no instructions for sanitizing or escaping the content read from the source files before it is written to the output file.
Audit Metadata