The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Inclusion of the blacklisted URL https://fitprofessionals.net within the directory catalog in references/directory-list.md and the template tracker in references/submission-tracker-template.csv. Providing malicious URLs as recommended resources increases the risk of user or agent interaction with malicious infrastructure.- [COMMAND_EXECUTION]: The skill instructs the execution of curl commands (curl -sIL [URL] | grep -i rel=) to verify the status of directory listings. This pattern can be exploited to perform Server-Side Request Forgery (SSRF) or facilitate information leakage if the agent or user interacts with untrusted or attacker-controlled domains included in the directory list.- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting untrusted external data without adequate security controls.
Ingestion points: The skill reads marketing context from files such as .agents/product-marketing.md and .claude/product-marketing.md to inform its plan.
Boundary markers: Absent; there are no delimiters or instructions to ignore embedded directives within the ingested marketing context.
Capability inventory: The skill suggests using shell commands (curl) and providing instructions to the agent based on ingested data.
Sanitization: Absent; the skill does not validate or sanitize the content of the marketing context files before processing.

directory-submissions