skills/dagster-io/erk/learned-docs/Gen Agent Trust Hub

learned-docs

Pass

Audited by Gen Agent Trust Hub on Feb 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions direct the agent to utilize standard CLI tools for documentation maintenance, specifically git mv for reorganizing files, grep for finding cross-references, and make fast-ci for validating markdown formatting and frontmatter integrity. It also uses a project-specific tool erk to synchronize generated index files from source metadata. These operations are consistent with the skill's primary purpose of documentation management within a development environment.\n- [PROMPT_INJECTION]: The skill implements a system of 'tripwires' and 'read_when' conditions defined in YAML frontmatter to influence agent behavior and navigation. This represents an indirect prompt injection surface where documentation content can trigger specific agent actions or warnings. \n
  • Ingestion points: Markdown documentation files located in docs/learned/.\n
  • Boundary markers: Uses standard YAML frontmatter delimiters (---).\n
  • Capability inventory: File relocation (git mv), searching (grep), and local script execution (make, erk).\n
  • Sanitization: The skill incorporates a validation step via make fast-ci to ensure frontmatter syntax and markdown formatting adhere to project standards.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 27, 2026, 09:34 AM