project-context
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it reads and processes untrusted external data (e.g., meeting transcripts, README files, repository documentation) and uses it to generate the foundational project context file.
- Ingestion points: According to the workflow in
SKILL.md(Step 2), the agent ingests data from README files, repository metadata, documentation indices, and meeting transcripts. - Boundary markers: The instructions do not specify any boundary markers (such as XML tags or explicit 'ignore instructions' delimiters) to wrap the external content during the ingestion process.
- Capability inventory: The skill possesses file-write capabilities (creating or updating
.agents/project-context.md) and is explicitly designed to serve as the 'foundational context' for other downstream skills, potentially spreading malicious instructions throughout the agent's workflow. - Sanitization: No sanitization, escaping, or validation steps are defined for the content extracted from external sources before it is interpolated into the context file.
Audit Metadata