The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/run_with_jira_env.py acts as a wrapper that executes arbitrary shell commands passed as arguments while injecting sensitive environment variables (such as JIRA_CLOUD_API_TOKEN) into the execution context. The agent is instructed to use this for Jira API interactions, but it provides a general-purpose execution bridge.
[DYNAMIC_EXECUTION]: The script scripts/check_jira_issue_mutation.py dynamically loads and executes the check_jira_text.py module at runtime using importlib.util. This is used for code sharing between local validation scripts.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from the Jira Cloud API (e.g., issue summaries, descriptions, and comments), which represents an untrusted external attack surface. Malicious instructions embedded in Jira content could potentially influence agent behavior.
Ingestion points: Jira issue data fetched via search and issue retrieval endpoints described in references/jira-cloud-api.md.
Boundary markers: The skill requires explicit user confirmation for all mutations and directs the agent to load references/safety.md at the start of the workflow.
Capability inventory: The skill can execute local scripts, git commands, and arbitrary shell commands with access to credentials via scripts/run_with_jira_env.py. It also performs network operations via curl.
Sanitization: The skill uses scripts/check_jira_text.py to validate Spanish text and filter out forbidden patterns like roadmap IDs or commit prefixes.

krt-jira-cloud-scribe