proposal-write
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill ingests untrusted user data like notes and review documents to modify proposal files, which is a surface for indirect prompt injection attacks.
- Ingestion points: User descriptions, seed references, and
<slug>-review.mdfiles used during the refinement process. - Boundary markers: The agent is instructed to prioritize
references/guidelines.mdas the authority, but no technical boundary or delimiter is applied to user-provided content. - Capability inventory: The skill is restricted to reading and writing Markdown files within the local workspace; it does not utilize network access or high-privilege shell commands.
- Sanitization: No explicit sanitization, escaping, or validation of user-provided notes is performed before they are incorporated into the documents.
Audit Metadata