reciprocity
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill consists entirely of Markdown documentation and educational examples. There are no scripts, binaries, or configuration files that could be executed by an agent.
- [SAFE]: The content is purely instructional, focusing on cognitive psychology (Regan's Coke Experiment, Cialdini's influence principles) and its application in business (AI freemium models, B2B sales). It encourages rigorous analysis of social obligations rather than bypassing safety or privacy protocols.
- [REMOTE_CODE_EXECUTION]: There are no patterns of remote code execution. External links point to informational websites (deciqai.com, doi.org) or standard version control (github.com) belonging to the vendor.
- [DATA_EXFILTRATION]: The skill does not access sensitive files or perform network operations. All analysis is performed within the user's provided context for coaching purposes.
- [PROMPT_INJECTION]: The skill contains no adversarial instructions meant to override agent behavior. The 'WAIT' tokens are used correctly within the skill's own logical flow for 'Coach mode' step-by-step interaction.
Audit Metadata