The Agent Skills Directory

[UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill instructs the agent to install optional companion tools using the command npx skills add mattpocock/skills --skill <name>. This involves the execution of the npx package runner to fetch and run code from a third-party GitHub repository (mattpocock/skills) which is not included in the verified trusted vendor list.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from various external and local sources, creating an attack surface for indirect prompt injection where malicious instructions could be embedded in data.
Ingestion points: The skill is instructed to read "Feature records", "product context", "external trackers", and the .waypoint/local.yaml configuration file.
Boundary markers: The instructions lack requirements for using delimiters or boundary markers to separate system instructions from the content ingested from these external files.
Capability inventory: The agent is granted the capability to read local filesystem contents and execute shell commands (via the npx installation instruction).
Sanitization: No sanitization, validation, or filtering of the external data is specified in the skill's instructions before the content is processed.

feature-spec