linkedin-teneo
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a function that handles external, untrusted data via a command-line interface, creating a potential vector for indirect prompt injection or input manipulation.
- Ingestion points: The
<url>input argument for theenrich_urlcommand inSKILL.mdallows arbitrary external web addresses to be provided to the agent. - Boundary markers: Absent; there are no instructions or delimiters defining boundaries to prevent the underlying model from interpreting hidden or malicious payloads within the referenced text or URL components.
- Capability inventory: The skill utilizes terminal shell execution via
~/teneo-skill/teneoto perform profile enrichment operations. - Sanitization: Absent; the skill configuration does not prescribe validation, escaping, or strict schema validation mechanisms for the parsed
<url>parameter prior to passing it to the local CLI environment.
Audit Metadata