The Agent Skills Directory

[DATA_EXPOSURE]: The skill attempts to inventory installed skills by accessing multiple hidden configuration and application directories in the user's home folder, including:
~/.codex/skills/ and ~/.agents/skills/
~/.claude/skills/
~/.cursor/skills/
~/.grok/skills/
~/.gemini/antigravity/skills/ and ~/.gemini/config/skills/
~/.gemini/antigravity-cli/skills/
[EXTERNAL_DOWNLOADS]: The skill instructs the agent to fetch information about uninstalled skills from an external website: https://skills.sh/. While this is used for informational purposes, it introduces a dependency on external content.
[INDIRECT_PROMPT_INJECTION]: The skill has a significant attack surface for indirect prompt injection:
Ingestion points: It reads local project manifest files (package.json, go.mod, pyproject.toml, Cargo.toml), a reference file from a separate skill (tek-skill/references/skills.md), and content from the external website https://skills.sh/.
Boundary markers: There are no specific instructions or delimiters provided to the agent to treat this ingested data as untrusted or to ignore instructions embedded within those files.
Capability inventory: The skill contains explicit negative constraints ("ห้ามรันคำสั่งติดตั้ง ห้ามแก้ไฟล์ ห้ามเปิดใบใด"
Do not run installation commands, do not edit files, do not open any items) which mitigates the risk of the agent acting on malicious data.
Sanitization: No sanitization or validation of the external content is mentioned.

tek-scan-skills