The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes trace data from external sources which can influence the agent's behavior.
Ingestion points: Trace spans, error messages, and attributes are retrieved from the Axiom platform using the axiom query tool (SKILL.md).
Boundary markers: The instructions do not establish clear delimiters or protective prompts to help the agent distinguish between data to be analyzed and instructions to be followed, which could allow malicious content in trace fields to manipulate the agent.
Capability inventory: The skill possesses the capability to execute axiom CLI commands and perform local file operations using Read, Grep, and Glob tools (SKILL.md).
Sanitization: There is no evidence of sanitization or validation of trace field contents before they are processed by the agent.

find-traces