The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs shell command execution to set up and interact with the project environment and external services.
Evidence: The skill executes local project setup and development scripts such as bin/setup, bin/dev, and uses configuration files like docker-compose.yml or Procfile (as seen in SKILL.md and references/running-the-code.md).
Evidence: Interaction with Jira is performed via the acli CLI tool to create and view work item comments (found in references/report-format.md).
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from external sources to define its operational criteria.
Ingestion points: The agent is instructed to read acceptance criteria, stakeholder comments from Jira ticket threads, and pull request descriptions or commit logs from git history (SKILL.md sections 2 and 5).
Boundary markers: The instructions lack specific delimiters or warnings to the agent to ignore potentially malicious instructions embedded within the ingested ticket data or git history.
Capability inventory: The agent possesses capabilities to execute shell commands (e.g., project scripts, acli), perform API requests, and drive browser automation (Chrome) (references/running-the-code.md).
Sanitization: There is no requirement or logic provided to sanitize or filter the external content before it is processed by the agent to determine its next steps.

qa-verification