{"id":19686,"plugin_id":"plugins_6a9bd8b75acc81919eb7c673307385b8","kind":"skill","collection_source":null,"comparison_source":null,"observed_at":"2026-09-30T23:15:43.123Z","digest":"2b718298e27ce0fe2128a580ce0284294b270885d148cce2e249313ddea53b7e","against":null,"payload":{"description":"Use when an important negotiation strategy, offer, meeting, or difficult conversation needs rehearsal against a credible resistant counterpart or independent stress-testing before execution.","included_files":[],"name":"roleplay-redteam","skill_md_contents":"---\nname: roleplay-redteam\ndescription: Use when an important negotiation strategy, offer, meeting, or difficult conversation needs rehearsal against a credible resistant counterpart or independent stress-testing before execution.\n---\n\n# Role-play and Red Team\n\n## Overview\n\nStress-test the current strategy using Freeman's role-play principle, with an AI isolation protocol that prevents simulated claims becoming facts. [F03]\n\n**Interface:** accept a versioned `SKILL_REQUEST`; return `SKILL_RESPONSE`. Never persist state.\n\n## Isolation rule\n\nThe supplied state snapshot is immutable during the exercise. Simulation output is **not evidence** about the real counterpart.\n\n## Method\n\n1. **Lock the snapshot.** Record `state_version`, current strategy, proposal, evidence and explicit assumptions.\n2. **Adopt a credible counterpart.** Model the strongest plausible resistance supported by the state; do not create a cartoon villain or easy opponent.\n3. **Stress-test.** Challenge weak assumptions, ambiguous terms, non-credible threats, exposed concessions and unattractive trades. Identify what a competent counterpart could ask for or attack.\n4. **Separate simulation from analysis.** If a role-play transcript is useful, return it only as an optional `deliverable.kind: roleplay`.\n5. **Debrief.** Return strongest counterarguments, weak assumptions, missing evidence, likely reactions and proposed tests/strategy changes. Simulated reactions remain hypotheses unless independently evidenced.\n6. **Reconcile externally.** The orchestrator classifies significant findings `ACCEPT`, `TEST`, `REJECT` or `DEFER`. F06 does not write them directly into canonical state.\n7. **Check staleness.** If the live state no longer matches `based_on_state_version`, warn rather than overwriting newer reasoning.\n\n## Safeguards\n\n- Attack the strategy, not the user personally.\n- Do not provide reassurance theatre; resistance should be meaningful and plausible.\n- Do not create new KNOWN facts from dialogue you generated.\n- Do not make the final negotiation recommendation or deal-acceptance decision; F08 owns independent acceptance.\n- Do not persist state or silently alter targets/BATNA.\n\n## Stop condition\n\nStop when further rounds no longer reveal material new vulnerabilities, evidence gaps or response problems.\n"},"changes":[],"summary":"First saved snapshot. No earlier version is available for comparison.","summary_kind":"deterministic","summary_metadata":{}}