The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/run_with_jira_env.py uses subprocess.run to execute shell commands provided as arguments after loading credentials from a local file. This is the primary mechanism for executing Jira API calls.
[COMMAND_EXECUTION]: The scripts scripts/setup_jira_env.py and scripts/check_jira_env.py invoke git via subprocess.run to manage and verify the exclusion of sensitive environment files from version control.
[EXTERNAL_DOWNLOADS]: The skill communicates with Atlassian Jira Cloud REST APIs (*.atlassian.net) to manage issues and metadata. These interactions are fundamental to the skill's purpose and target a well-known service.
[PROMPT_INJECTION]: The skill processes data from an external source (Jira issues) which could contain malicious instructions. It incorporates a mandatory evidence chain for indirect prompt injection:
Ingestion points: Jira issue search results and details processed in SKILL.md (Workflow steps 2, 3, and 4).
Boundary markers: Explicitly defined in references/safety.md and the 'Required Confirmations' section of SKILL.md, requiring user approval before any mutation.
Capability inventory: Subprocess execution in scripts/run_with_jira_env.py and file writing in scripts/setup_jira_env.py.
Sanitization: scripts/check_jira_text.py performs regex-based validation of Spanish semantic text to block operational patterns like commit prefixes and PR chatter.

krt-jira-cloud-scribe