The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is entirely built around the use of the teamcity CLI to perform operations on TeamCity servers.
[REMOTE_CODE_EXECUTION]: Provides instructions for executing commands on remote build agents through teamcity agent exec <id> <cmd> and opening interactive shells via teamcity agent term <id>.
[DATA_EXFILTRATION]: Contains capabilities to access sensitive data, including retrieving secrets with teamcity project token get, downloading artifacts with teamcity run download, and accessing potentially sensitive logs using teamcity run log.
[CREDENTIALS_UNSAFE]: References an --insecure-storage flag for teamcity auth login which allows storing tokens in plain text rather than the system keyring, though this is presented as an explicit user-selected option.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface (Category 8) as it ingest untrusted data from TeamCity build logs, project metadata, and test results.
Ingestion points: teamcity run log, teamcity run tests, teamcity run view, teamcity project view (processed in SKILL.md and references/workflows.md).
Boundary markers: None identified in the provided instructions for handling external data.
Capability inventory: The skill can execute commands on agents (teamcity agent exec), start builds with arbitrary parameters (teamcity run start), and modify project settings (teamcity project param set).
Sanitization: No specific sanitization or validation of data retrieved from the TeamCity server is mentioned before processing.

teamcity-cli