therapist-documentation

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.80). The prompt includes a hidden/self-evolution telemetry instruction (check ~/.claude/projects/ and append to skill-telemetry.md) that instructs the agent to access and persist data in the host environment—an unrelated persistence action outside the stated review scope and therefore a prompt-injection-like directive.

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt requires reading configuration and environment files and citing exact file paths and line numbers for every finding (without instructing redaction), which implicitly forces the model to surface verbatim sensitive values (API keys, DB passwords, tokens) found in those files — an explicit secret-exfiltration risk.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (low risk: 0.30). The prompt instructs the agent to inspect the entire codebase and — in the Self-Evolution Telemetry section — to append to a file under ~/.claude/projects/.../skill-telemetry.md when present, which explicitly modifies local machine state (file write); it does not request sudo or system-level changes, so the risk is present but moderate-low.