The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill includes a reference to the domain 'fitprofessionals.net' (Fit Pro Directory) in both 'references/directory-list.md' and 'references/submission-tracker-template.csv'. This domain is currently blacklisted and identified as malicious by security scanners.
[COMMAND_EXECUTION]: In 'SKILL.md', the instructions direct the agent to verify backlink status using a shell command: 'curl -sIL https://directory.com/your-listing | grep -i rel='. Executing shell commands against untrusted URLs provided by users or external sites (via the tracker) is a security risk that could lead to command injection if input is not strictly validated.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from two primary sources:
Ingestion Points: Reading product marketing context from '.agents/product-marketing.md' and fetching content from external directory URLs during submission verification.
Capability Inventory: The agent is equipped with 'curl' for network operations as described in 'SKILL.md'.
Boundary Markers: No specific boundary markers or 'ignore embedded instructions' warnings are implemented to protect against malicious instructions hidden in external marketing files or directory pages.
Sanitization: There is no mention of escaping or validating the content retrieved from external sources before it is used by the agent.
[METADATA_POISONING]: The skill instructions ('SKILL.md') and reference files contain deceptive context, making multiple references to '2025' and '2026' algorithms and market conditions (e.g., '2026 PH algorithm', 'G2 Summer 2025 paid plans'). This hallucinated future-dating can mislead the agent and users about the accuracy and timeliness of the security and operational advice provided.

directory-submissions