The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructions in SKILL.md and the fetch_conditions.py script utilize uvx to execute code directly from a remote Git repository (github.com/dreamiurg/peakbagger-cli.git). This involves downloading and running an external package at runtime.\n- [COMMAND_EXECUTION]: The Python tools fetch_conditions.py and cloudscrape.py execute shell commands via subprocess.run. fetch_conditions.py invokes the PeakBagger CLI tool, while cloudscrape.py runs a command to install the Chromium browser dependency.\n- [EXTERNAL_DOWNLOADS]: The skill performs several external requests to fetch conditions and geographic data from the Open-Meteo API, the FCC Area API, and OpenStreetMap Overpass. It also triggers the download of the Chromium browser binary via the patchright library during its first execution.\n- [INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection because it fetches and processes untrusted data from multiple external websites (e.g., trip reports from WTA, AllTrails, and SummitPost) to generate summarized mountaineering reports.\n
Ingestion points: Web content from external mountaineering domains is fetched and processed by sub-agents as described in SKILL.md.\n
Boundary markers: The skill includes a prominent safety disclaimer in its report template (assets/report-template.md).\n
Capability inventory: The skill possesses file-writing, network access, and command execution capabilities.\n
Sanitization: No specific sanitization or filtering of the scraped web content is described before it is synthesized into the final report.

route-researcher