night-market-docs-and-writing
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute several local maintenance and validation scripts. These include Python utilities for checking markdown links (
check-markdown-links.py), validating skill descriptions (validate_budget.py), and fixing formatting (fix_descriptions.py). It also utilizes standard shell utilities likegrep,ls, andrgfor repository inspection andmakefor documentation synchronization. - [INDIRECT_PROMPT_INJECTION]: The skill operates on a wide range of repository documentation files, creating an attack surface where malicious content in processed files could influence agent behavior.
- Ingestion points: The skill reads and writes to numerous files including
docs/*.md,book/src/*.md,CHANGELOG.md, and ADRs indocs/adr/. - Boundary markers: The instructions do not specify the use of delimiters or boundary markers when reading or interpolating file content into prompts.
- Capability inventory: The skill has access to the local filesystem for reading and writing, can execute repository-local Python and shell scripts, and refers to sibling skills for interacting with GitHub Discussions via GraphQL.
- Sanitization: The skill implements a robust 'slop-check' system consisting of three layers (P0 critical hits, document economy scoring, and vocabulary tiers) designed to identify and block machine-generated prose patterns, identity leaks, and hallucinated identifiers.
Audit Metadata