The Agent Skills Directory

[COMMAND_EXECUTION]: The skill reads project files such as AGENTS.md and STRATEGY.md and performs repository searches to establish context for brainstorming tasks.
[DATA_EXFILTRATION]: The skill supports uploading generated requirements to the Proof editor for review and can research context via Slack tools. These operations are consistent with the skill's stated purpose and integration within its vendor ecosystem.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it ingests user input and project files to generate documents.
Ingestion points: User-provided arguments and repository files including AGENTS.md, STRATEGY.md, and existing documents.
Boundary markers: User-provided feature descriptions are delimited using <feature_description> XML tags.
Capability inventory: File system read/write operations for markdown documents and delegation to other tools in the ecosystem.
Sanitization: The skill does not describe explicit sanitization of ingested content before interpolation.

ce-brainstorm