get-context
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted user input from
$ARGUMENTSand local dependency manifests (package.json,go.mod,pyproject.toml,Cargo.toml) and interpolates the data verbatim into a prompt for a sub-agent. This creates a surface for both direct and indirect prompt injection attacks. - Ingestion points: Command-line arguments and local project configuration files.
- Boundary markers: None; the instructions explicitly pass targets verbatim to the sub-agent without delimiters or ignore-instructions warnings.
- Capability inventory: Repository cloning (
git clone) and web content fetching. - Sanitization: No validation or escaping is performed on the ingested strings before they are passed to the agent.
- [EXTERNAL_DOWNLOADS]: The workflow includes
git cloneas a primary method for acquiring context from remote repositories, which involves downloading external code. - [COMMAND_EXECUTION]: The skill logic classifies input strings to trigger repository cloning or external lookup tools based on user-supplied or file-derived data.
Audit Metadata