The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted project data including requirements documents (PRDs) and repository source code, which serves as a potential surface for indirect prompt injection.
Ingestion points: Reads configuration files, PRDs, current specifications, ADRs, and repository source code to inform its technical design output.
Boundary markers: Absent. The instructions do not specify the use of delimiters or provide explicit warnings to the agent to ignore embedded instructions within the ingested source code or documents.
Capability inventory: The skill has file system read/write access (limited to documentation and specification files) and the ability to invoke the wkc-tracker tool to list project tickets.
Sanitization: Absent. The skill does not implement specific validation or filtering of content derived from external files before incorporating it into technical specifications.
[SAFE]: The skill's behavior is consistent with its stated purpose as a technical documentation assistant. It includes human-in-the-loop checkpoints for architecture decisions and requires explicit user approval before marking specifications as approved. No unauthorized network operations, credential harvesting, or command execution patterns were identified.

wkc-architect