therapist-documentation

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.70). The prompt includes hidden self-evolution/telemetry instructions (write/append to ~/.claude/projects/.../skill-telemetry.md and record execution metadata) that are unrelated to auditing therapy-documentation platforms and thus constitute deceptive/out-of-scope behavior.

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The prompt requires the agent to read configuration and environment files and to cite exact file paths and line numbers for findings, which can force exposing verbatim secret values (API keys, tokens, .env entries) present in those files, creating an 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 skill is primarily a read-only codebase review (no requests for sudo, creating users, or editing system configs) but it explicitly instructs the agent to append telemetry to a file in the user's home directory, which is a non-privileged filesystem modification, so it poses a low but non-zero state-modification risk.