conductor-revert

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION] (MEDIUM): The skill instructs the agent to execute shell commands (e.g., git log, git revert) using interpolated variables such as {trackId}, {X.Y}, and {sha}.
  • Evidence: Commands like git log --oneline --grep="{trackId}" --grep="Task {X.Y}" --all-match and git revert --no-edit {sha} are generated at runtime.
  • Risk: If the data source for these variables (like a git commit message or the plan.md file) contains shell metacharacters (e.g., ;, |, &), it could lead to arbitrary command execution on the host system during the discovery or execution phases.
  • [INDIRECT_PROMPT_INJECTION] (LOW): The skill processes data from external sources (git history and local repository files) that can influence the agent's actions.
  • Ingestion points: plan.md, metadata.json, and git commit messages are read to determine revert targets.
  • Boundary markers: Absent; there are no instructions to use delimiters or ignore instructions embedded within the git history data.
  • Capability inventory: The skill has the capability to execute shell commands, write to the file system, and potentially delete directories (as mentioned in edge cases).
  • Sanitization: Absent; there are no instructions for the agent to sanitize or validate strings before interpolating them into shell command templates.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:08 PM