skills/dagster-io/erk/rename-swarm/Gen Agent Trust Hub

rename-swarm

Pass

Audited by Gen Agent Trust Hub on Apr 8, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface because it ingest untrusted data from local files and processes it using sub-agents.
  • Ingestion points: Local files discovered via the Grep tool as described in Step 1 of SKILL.md.
  • Boundary markers: The sub-agent prompt template includes a {boundary_constraints} variable for exclusions, but lacks formal delimiters to prevent the agent from following instructions embedded within the files being edited.
  • Capability inventory: The skill uses the Task agent to launch sub-agents with the Edit tool for file modification. The main workflow uses Grep for file discovery and devrun for executing verification commands.
  • Sanitization: No sanitization or validation of file content is performed before it is passed to the sub-agents for editing.
  • [COMMAND_EXECUTION]: The workflow incorporates the execution of search tools and testing agents to verify changes.
  • Evidence: Step 1 utilizes the Grep tool to identify target files. Step 5 uses the devrun agent to execute shell commands for running CI tests, type checking, and linting.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 8, 2026, 03:21 AM