The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted data in the form of high-level requirements or feature descriptions to generate ServiceNow records. There is a potential surface for indirect prompt injection where malicious instructions embedded in a requirement description could influence the agent's behavior during the decomposition process.
Ingestion points: Feature descriptions, stakeholder requests, and high-level requirements (SKILL.md, Procedure Section).
Boundary markers: Not explicitly defined in the prompt templates provided in the instructions.
Capability inventory: SN-Create-Record, SN-Update-Record, and REST API POST/PATCH operations on ServiceNow tables.
Sanitization: No explicit sanitization or validation of the input requirements is specified before interpolation into story records.
[COMMAND_EXECUTION]: The skill lists Bash as a native tool, which is utilized for executing curl commands against the ServiceNow Table API. The commands used are strictly limited to standard record lifecycle operations (GET, POST, PATCH) and do not involve executing external scripts or piping network data to a shell.

agile-story-generation