The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple shell commands to perform its core functions.
Uses gh issue create to interact with GitHub.
Uses jira-cli to interact with Jira.
Performs file system operations including directory renaming (mv).
Executes git branch renaming commands.
[CREDENTIALS_UNSAFE]: The skill is designed to access environment variables for configuration.
Specifically looks for JIRA_PROJECT and JIRA_DOMAIN to configure the issue creation process.
[DATA_EXFILTRATION]: Local workflow data is transmitted to external platforms.
Contents from 01_KICKOFF.md (Title, Goal, Success Criteria) are sent to GitHub or Jira as issue descriptions.
Local configuration and state are used to determine destination endpoints.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through local file content.
Ingestion points: Data is read from 01_KICKOFF.md and state.json (specifically the 'Goal' and 'Success Criteria' sections).
Boundary markers: None. Content is extracted and used directly in command-line arguments.
Capability inventory: The skill has the ability to read/write local files, rename directories/branches, and execute commands via the gh and jira-cli tools.
Sanitization: There is no evidence of sanitization or escaping of the content extracted from markdown files before it is passed as a string to the shell commands (e.g., as the body of a GitHub issue).

wf0-promote