The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands to manage the project environment. Evidence includes the use of echo to modify the .gitignore file and git commands (add, commit) to stage and record changes to the local repository. These operations are performed automatically as part of the plan generation workflow.- [DATA_EXFILTRATION]: The skill performs network operations via git push to upload the generated implementation plan and modified project files to a remote repository. This constitutes a data transit point where local workspace content is sent to an external server, which is the intended behavior for team collaboration.- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. * Ingestion points: Processes user-provided "spec or requirements" as described in the skill header. * Boundary markers: None present; the skill does not use delimiters to isolate the untrusted input from its own instructions. * Capability inventory: Performs file writes, git commit, and git push (documented in SKILL.md). * Sanitization: None present; the skill does not validate or filter the content of the input spec before using it to generate the plan that is subsequently committed and pushed.

team-plan