The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates shell command execution on local and remote systems through the teamcity agent exec and teamcity agent term commands, and interacts with the TeamCity server via its CLI.
[DATA_EXFILTRATION]: Provides functionality to retrieve stored secrets and tokens using teamcity project token put/get and download build artifacts with teamcity run download.
[REMOTE_CODE_EXECUTION]: Includes native support for executing arbitrary code on TeamCity build agents through the CLI's agent management subcommands.
[PROMPT_INJECTION]: Identified an indirect prompt injection surface in the babysit-build autonomous agent.
Ingestion points: The agent reads raw build logs (teamcity run log --raw) and test outputs (teamcity run tests) to diagnose failures.
Boundary markers: No specific delimiters or instructions to ignore embedded commands in logs are defined.
Capability inventory: The agent has Bash, Write, and Agent permissions, enabling it to modify the repository and execute commands on agents to apply fixes.
Sanitization: No evidence of sanitization or validation of the log content before it influences agent decisions.

teamcity-cli