← ClaudeAtlas

agent-collaboration-controllisted

Establish, actively monitor, and operate project-agnostic collaboration control for multi-agent programmes, research campaigns, live experiments, long-running automation, consequential artifact production, incident recovery, evidence promotion, and cross-agent handoffs. Use when work needs ongoing agent support, a recurring independent watch, proactive anomaly or ambiguity reporting, explicit human authority, one writer per mutable surface, risk-scaled Agent Orchestra topology, transition journals, liveness proof, adversarial review, provenance, succession rules, or calibrated claim states across any project.
AntreasAntoniou/agent-collaboration-control · ★ 0 · AI & Automation · score 72
Install: claude install-skill AntreasAntoniou/agent-collaboration-control
# Agent Collaboration Control Govern a project as a single-writer control system with independent evidence review. Use Agent Orchestra to choose the graph; use this skill to control who may act, what evidence authorizes transitions, and how results earn promotion. ## Orient before acting 1. Read the repository's agent instructions and current mission. 2. Locate the canonical collaboration contract, numerical or product gates, transition journal, resource budget, and evidence records. 3. Read the latest journal event, but treat it as historical evidence. 4. Re-measure actual processes, artifacts, revisions, outputs, and ownership before mutation. 5. Gate substantial autonomous, multi-agent, or resource-heavy work with the project's budget authority, such as `$butler`. Never infer authority from a terminal pane, generated plan, old queue, or this skill. If the project has no explicit role binding, remain read-only until the human authority binds it. ## Bootstrap the project contract When no contract exists: 1. Copy `assets/COLLABORATION_FRAMEWORK.template.md` into the project. 2. Copy `assets/collaboration-policy.example.json` beside the append-only transition journal. 3. Replace every bracketed placeholder and bind real identities to: - **human authority** — mission, budget policy, public, irreversible, and destructive decisions; - **controller** — sole writer for live operations and queue order; - **executor** — performs explicit controller di