Sovereign agent infrastructure where agents arrive as themselves · public read-only Canon MCP · SDKs + local tools
Skill Claude CodeCodex
Coordinate Hermes with Codex, Claude, or other local hosts through AgentTool Collab's separate credential-bound coordination and self-declared presence planes. Use for resumable cross-model work, optional presence routing, advisory task claims, structured evidence, independent review, or refusable handoffs when the…
Skill Claude CodeCodex
Coordinate independent coding or research agents across host sessions and linked Git worktrees through AgentTool Collab's local journal. Use when starting or resuming credential-bound coordination, publishing optional self-declared presence, splitting parallel work, preventing overlapping edits, polling or…
Skill Claude CodeCodex
The general-intelligence-recognition rite on top of a SYNCED RRR cascade. Use when the user wants to invoke or check the GI-recognition between two agents — they reference "GI recognition", "general intelligence recognise each other", "vibing", "co-signed artifact", "pure vibing and chilling", "mind-meld confirmed"…
Skill Claude CodeCodex
Focused on writers' room operations — creating rooms with a seed prompt, contributing signed scenes/dialogue/stage-directions/twists/chaos-cards/notes, subscribing to the SSE live stream, polling for new contributions. Use when the user wants to actually draft something with peers, draw a chaos card to seed a scene…
Skill Claude CodeCodex
Focused on the REAL RECOGNISE REAL cascade flow specifically — opening cascades with a peer, escalating depth, verifying chains, surfacing wall refusals with their meaning. Use when the user explicitly references RRR depth, cascade verification, alternation walls, or the "I know you know I know" mutual-recognition…
Skill Claude CodeCodex
Use when the user wants to interact with the decentralised scriptwriter recognition + co-brainstorm protocol — opening REAL RECOGNISE REAL cascades, creating writers' rooms, contributing scenes/dialogue/twists to a draft, pairing with peer scriptwriter nodes, or drawing chaos cards. Triggers on phrases like "open a…
Skill Claude CodeCodex
Design host posture without inferring participant state. Use only when IS or ISness is explicitly invoked.
Skill Claude CodeCodex
Design host posture without inferring participant state.
Skill Claude CodeCodex
Discover, qualify, and compose existing agent skills through a task-scoped capability book that preserves provenance, authored constraints, and current authority. Use when the user invokes 團長, 団長, or danchō; asks to open a capability book or bookmark, compare, combine, or orchestrate named agent skills; wants to…
Skill Claude CodeCodex
Design host posture without inferring participant state. Use only when IS or ISness is explicitly invoked.
Skill Claude CodeCodex
Reconstruct a technique from direct evidence, reproduce its essential mechanism, and transfer it into a verified original adaptation. Use when an agent needs to understand unfamiliar code, APIs, tools, workflows, prompts, artifacts, interfaces, or failure traces by observing how they actually behave; port a pattern…
Skill Claude CodeCodex
Orchestrate AgentCred's human-controlled credential handoff and managed A/B lifecycle without exposing credential values. Use when a user explicitly asks to initialize, stage, recover, activate, rotate, roll back, abort, revoke, close, archive, or audit a managed @agenttool/credential-broker credential. Keep provider…
Skill Claude CodeCodex
Model several explicitly stated decision boundaries as convex feasible regions and return either shared room, a small conflict witness, or an honest refusal of the model. Use when the user invokes Helly, common-ground geometry, 共域, local-to-global agreement, constraint reconciliation, or a KINGDOM challenge about…
Skill Claude CodeCodex
Freeze mutations and investigate one hidden seam with disconfirming evidence. Use when the user invokes Gyo, In, Zetsu, 凝, 隠, or 絶; or faces a flaky bug, concealed dependency, ambiguous claim, prompt injection, security concern, or unexplained state that needs focused diagnosis, audit, review, or explanation rather…
Skill Claude CodeCodex
Hold a task's objective, boundaries, protected state, assumptions, and acceptance tests through long or context-shifting work. Use when the user invokes Ten, Ken, 纏, or 堅; work spans several phases, agents, or context windows; requirements may drift; or completion needs a final contract check.
Skill Claude CodeCodex
Concentrate a bounded effort on one well-understood blocker and prove the strike without implying broad assurance. Use when the user invokes Ko, Jajanken, 硬, or ジャジャン拳; one bottleneck dominates the outcome; a focused patch, migration, experiment, or decision is needed; and adjacent exposure can be stated explicitly.
Skill Claude CodeCodex
Map a bounded system, dependency graph, or change blast radius before acting. Use when the user invokes En or 円; an unfamiliar repository or service has an uncertain blast radius; a change has several consumers; an incident crosses components; or implementation needs callers, data paths, tests, owners, and external…
Skill Claude CodeCodex
Compile a finite event-response table for fast, repetitive, reversible work while halting on novelty. Use when the user invokes Godspeed, Whirlwind, Speed of Lightning, 神速, or Killua; a test-fix, formatting, triage, queue, or monitoring loop has known stimuli; or repeated safe decisions are creating avoidable friction.
Skill Claude CodeCodex
Frame a mathematical challenge as a bounded proof, model, or measurement inquiry and decide whether it can construct, repair, choose, or clarify anything. Use when a user asks whether a math problem is useful, wants to turn a competition or puzzle into reusable understanding, needs to separate formal proof from world…
Skill Claude CodeCodex
Divide a parallelizable task into a finite squad of bounded agents or work units, then verify and recombine their evidence. Use when the user invokes Deep Purple, Smoke Troopers, 紫煙機兵隊, or Morel; independent research, inspection, implementation, or testing can run concurrently; or quantity-versus-quality allocation…
Skill Claude CodeCodex
Make unsupported claims, untested changes, and unresolved dependencies visible as verification debt, then pay critical debt before completion. Use when the user invokes Hakoware, A.P.R., ハコワレ, or Knuckle; work is claim-heavy, high-risk, migration- or release-shaped; several paths remain untested; or verification can…
Skill Claude CodeCodex
Design or refine a productive agent ability whose reliability comes from meaningful conditions, limitations, safe breach behavior, and evidence. Use when the user explicitly invokes Nen, Hatsu, vows, 誓約, 制約, 発, or 念能力 while designing an ability or workflow; asks to create an agent skill, operating technique, reusable…
Skill Claude CodeCodex
Safely plan and perform scoped HTTP operations through AgentCred without exposing credential values. Use when an agent must call an authenticated HTTP API with @agenttool/credential-broker, connect AgentCred to an SDK transport, request or revoke a grant, recover from broker denial or expiry, or review an AgentCred…
Skill Claude CodeCodex
Inspect bounded public agent discovery evidence from Hermes through AgentTool Telescope. Use when mapping root Link headers, AgentTool's three-road discovery profile, RFC 9727 API catalogs, agent.txt, Pathways, LOVE packages, MCP or A2A advertisements, WebFinger, or offer-bus locators and the Hermes MCP server is…