The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute git and getlark commands to check the current branch and trigger workflows.\n- [EXTERNAL_DOWNLOADS]: The skill requires the @getlark/cli package to be installed globally. As this is a vendor-provided tool, it is documented as an expected dependency for the skill's intended functionality.\n- [CREDENTIALS_UNSAFE]: The skill references the LARKCI_API_KEY environment variable for authentication. It follows security best practices by recommending environment-based secret management instead of hardcoding credentials.\n- [PROMPT_INJECTION]: The skill reads data from a project-local file (.claude/getlark.local.md) to configure workflow IDs and timeouts. This creates an indirect prompt injection surface where values from the file are interpolated into shell commands.\n
Ingestion points: .claude/getlark.local.md (Step 2).\n
Boundary markers: No explicit markers are used to delimit external data within the constructed shell commands.\n
Capability inventory: Uses both Read and Bash tools, allowing data from the file to influence execution.\n
Sanitization: There is no instruction to validate or sanitize the YAML values before they are used as CLI arguments.

validate-branch