analyzing-recent-project-state
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes a set of predefined, read-only Git commands to collect evidence. It also executes an internal shell script,
scripts/validate-output.sh, to validate the structure of its outputs. These executions are limited to safe, local operations. - [INDIRECT_PROMPT_INJECTION]: The skill processes external data from Git repositories (commit messages, file contents). It addresses this risk through strict operating posture instructions that mandate treating all retrieved content as data for summarization and explicitly forbids interpreting it as new instructions.
- Ingestion points: Repository path, Git history, and file contents accessed via subagents.
- Boundary markers: Explicit instructions in
SKILL.mdand subagent markdown files to ignore instructions embedded in retrieved data. - Capability inventory: Read-only Git commands, file system reads, and execution of the included validation script.
- Sanitization: A deterministic
awk-based script validates all handoff payloads to prevent malformed or malicious data from influencing the workflow.
Audit Metadata