The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves fetching and analyzing external web content via scripts/fetch_page.py. This introduces a surface for indirect prompt injection, where an attacker-controlled website could include malicious instructions meant to subvert the agent's logic.\n
Ingestion points: scripts/fetch_page.py (external URLs), SKILL.md (connectors and data sources).\n
Boundary markers: SKILL.md and references/skill-contract.md contain explicit instructions for the agent to treat scraped material as untrusted input and ignore embedded instructions.\n
Capability inventory: The skill includes scripts for file writing (scripts/generate_markdown_report.py, scripts/generate_pdf_report.py, scripts/schema_generator.py), health scoring, and fetching web data.\n
Sanitization: Basic HTML unescaping and regex extraction are applied in scripts/fetch_page.py, but the skill relies primarily on instructional guardrails to prevent exploitation.\n- [EXTERNAL_DOWNLOADS]: The script scripts/generate_pdf_report.py references an external Python library, reportlab, and provides a command for the user to install it manually. There is no automated or silent execution of remote code.\n
Evidence: scripts/generate_pdf_report.py mentions python3 -m pip install reportlab in a user-facing error message.

seo-backlinks