hermes-paperclip-adapter

Pass

Audited by Gen Agent Trust Hub on Sep 12, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructions direct the user to install external software packages, including hermes-agent via pip and hermes-paperclip-adapter via npm, from public registries.
  • [COMMAND_EXECUTION]: The adapter invokes the hermes CLI tool and enables toolsets like terminal and file, which allow for arbitrary command execution and file manipulation on the host system as part of its primary function.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data which could contain malicious instructions designed to manipulate the AI agent's behavior.
  • Ingestion points: The promptTemplate in SKILL.md interpolates external data from Paperclip issues and comments via variables such as {{taskBody}}, {{taskTitle}}, and {{commentId}}.
  • Boundary markers: None. The template uses logical tags for structure but does not include explicit delimiters or instructions to the LLM to ignore potentially malicious embedded commands in the task content.
  • Capability inventory: The skill documentation lists toolsets such as terminal, file, web, browser, and code_execution, providing the agent with broad system and network access.
  • Sanitization: No evidence of input sanitization or filtering is described in the integration instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 12, 2026, 06:50 PM
Security Audit — agent-trust-hub — hermes-paperclip-adapter