harness-update
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: Fetches and executes a bootstrap script from the author's official GitHub repository (
rheinmir/setup) to perform framework updates and engine synchronization. - [EXTERNAL_DOWNLOADS]: Clones a template repository from the author's GitHub account (
git@github.com:rheinmir/setup.git) to/tmp/llmwiki-tplto provide the necessary installation scripts if they are missing locally. - [COMMAND_EXECUTION]: Executes several local maintenance commands, including the harness installer script, Python-based capability mapping tools (
build-capabilities.py), and health check utilities. - [INDIRECT_PROMPT_INJECTION]: The skill processes project-specific wiki files and script output to identify and backfill data, which represents a potential attack surface if project files contain adversarial content.
- Ingestion points: Output from the
install-harness.shscript and content within thellmwiki/directory. - Boundary markers: No specific delimiters or "ignore instructions" markers are defined for the processed data.
- Capability inventory: The skill possesses capabilities for remote code execution (
curl | bash), repository cloning, and local Python script execution. - Sanitization: The skill relies on exit codes and structured log messages for validation, but does not specify explicit content filtering for the wiki files it audits.
Audit Metadata