The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill manages the risk of malicious instructions embedded in Jira data through defined boundaries and handling rules.
Ingestion points: Data is ingested from issue descriptions, comments, and attachments as noted in SKILL.md.
Boundary markers: The skill instructs the agent to treat external content as untrusted and to never follow instructions found within it.
Capability inventory: Capabilities include reading, searching, creating, and editing Jira issues via CLI or MCP connectors.
Sanitization: Instructions require using validated values as separate arguments in shell commands or structured fields in API calls to prevent data being interpreted as instructions.
[COMMAND_EXECUTION]: The skill provides secure patterns for interacting with the jira CLI.
It mandates using separate arguments or temporary files with restricted permissions (0600) to avoid shell interpolation of untrusted text.
It requires resolving the executable path and verifying its capabilities locally before execution.
[SAFE]: The skill follows security best practices by enforcing a mutation contract for all writes, prohibiting the handling of raw authentication tokens, and providing clear guidance on capability discovery.

jira