explore
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill employs dynamic context injection (the '!' syntax) to execute local git commands like 'git branch', 'git status', and 'git rev-parse' at load time. These commands are hardcoded and used for environment awareness, presenting minimal injection risk.
- [DATA_EXFILTRATION]: Instructions direct the agent to analyze project configuration and local settings, which may include sensitive paths such as '.env' files. While the skill provides a behavioral mitigation by instructing the agent to 'verify presence' of tokens without reading their values, the inherent capability to access these files exists.
- [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because its core function is to ingest and process untrusted external data (source code, git history, and documentation).
- Ingestion points: Reads files and logs via Glob, Grep, and Read tools.
- Boundary markers: The skill does not instruct the agent to use specific delimiters or 'ignore-instructions' blocks when processing codebase data.
- Capability inventory: The agent can execute shell commands (bash) and write artifacts to a local memory directory.
- Sanitization: No sanitization or content filtering is performed on the ingested code or file contents.
Audit Metadata