heurist-mesh-skill

Pass

Audited by Gen Agent Trust Hub on Jul 23, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the installation of the Foundry toolkit via a piped shell command (curl -L https://foundry.paradigm.xyz | bash). This download originates from a well-known service in the blockchain development community and is used for local cryptographic signing of payment authorizations.
  • [COMMAND_EXECUTION]: The agent is directed to use CLI tools including curl for API communication and cast (part of Foundry) for signing EIP-712 messages. These operations are restricted to the vendor's documented endpoints (mesh.heurist.xyz and mesh.heurist.ai).
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it processes untrusted data from external sources such as Twitter/X timelines and web scrapes via the TwitterIntelligenceAgent and ExaSearchDigestAgent.
  • Ingestion points: Tools such as user_timeline, twitter_search, and exa_scrape_url in SKILL.md.
  • Boundary markers: None explicitly defined in the instructions for isolating untrusted tool output.
  • Capability inventory: The agent has network access via curl to interact with the Heurist API.
  • Sanitization: No specific sanitization or filtering logic is described for the content retrieved from social media or web searches.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 23, 2026, 04:05 PM