trail-distil

Pass

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes shell commands (grep, ls, sort, uniq) to analyze file content within the .trail/entries/ directory and to locate existing rule files like CLAUDE.md or .cursorrules. These commands are used for data analysis and do not involve network connectivity or privilege escalation.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from files in the .trail/ directory, which could be used as a vector for prompt injection if the source files contain malicious instructions disguised as corrections.
  • Ingestion points: Reads all files within the .trail/entries/ directory and existing project configuration files (e.g., CLAUDE.md, .cursorrules).
  • Boundary markers: The instructions do not define clear delimiters or instructions for the agent to ignore potential commands embedded within the analyzed corrections.
  • Capability inventory: The skill has the ability to read files, execute local shell commands for analysis, and propose modifications to critical project configuration files.
  • Sanitization: No explicit sanitization or validation of the distilled content is performed before the agent processes it or proposes it as a new rule.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 27, 2026, 09:48 AM
Security Audit — agent-trust-hub — trail-distil