harness-collaboration

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly provides a bearer token in the prompt and requires the agent to include that token verbatim in the Authorization header of callback POSTs, which forces the LLM to handle/output a secret value directly (exfiltration risk).

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). The GitHub link points to a third‑party repository (BankrBot/skills) being used as the install source — a common vector for unvetted/executable code from an unknown account and therefore potentially high risk, while tryharness.ai is the official allowlisted callback host and not suspicious.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill is explicitly designed to execute real financial actions using a Harness-provisioned Bankr wallet. It states the agent "execute[s] with the funds in that provisioned wallet," references an action class financial_onchain, and requires proposals/one-use authorizations with explicit caps, chains, tokens, venues, and amounts. These are specific crypto/transaction execution capabilities (wallet/on-chain actions), not generic tooling, so it grants direct financial execution authority.