workspace-fit

Pass

Audited by Gen Agent Trust Hub on Aug 4, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to utilize a specific set of read-only shell commands for repository inspection, including pwd, git status, git ls-files, find, ls, file, sed, rg, and git grep. These commands are used to gather evidence without modifying the repository state.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) as it processes data from external, potentially untrusted repository files.
  • Ingestion points: The workflow involves reading orientation files like README and CLAUDE.md, as well as manifests, lockfiles, and configuration scripts (SKILL.md).
  • Boundary markers: The skill does not define technical delimiters or specific parsing rules to isolate content read from files from the agent's instructions.
  • Capability inventory: The agent is authorized to use shell utilities for directory listing, file identification, and content searching (ls, find, grep).
  • Sanitization: While the instructions forbid reading secrets or running setup scripts, there is no technical sanitization or validation of the text content ingested from the repository files.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 4, 2026, 09:03 PM
Security Audit — agent-trust-hub — workspace-fit