@llblab/pi-kit 0.1.0

package/node_modules/@llblab/pi-actors/dist/skills/recipe-memory/recipes/tool-review.json

@@ -0,0 +1,14 @@
+{
+  "async": true,
+  "args": [
+    "input_path:path",
+    "model:string",
+    "thinking:string"
+  ],
+  "defaults": {
+    "model": "{current_model}",
+    "thinking": "{current_thinking}"
+  },
+  "description": "Read-only background reviewer for one immutable 36-tool evolution portfolio.",
+  "template": "pi -p --model {model} --thinking {thinking} --no-tools @{input_path} Review the attached immutable pi-actors active-tool portfolio. Evaluate all 36 tools independently using lifetime and revision usage, adaptability, redundancy, safety, contract quality, and likely future usefulness. Treat source/name and sha256 fields as batch-local opaque occurrence/content-group identifiers, return them verbatim, and infer exact duplication only from a shared content group. Choose exactly one keep, evolve, demote, or merge decision per tool; there is no action quota. evolve may only rename one unchanged captured recipe, demote moves the unchanged recipe to draft memory, and merge may only deduplicate at least two canonically identical captured recipes. Never return recipe or outputs fields; replace and split require explicit operator-authored mutation and are unavailable to automatic review. Do not mutate, register, move, or delete recipes. End stdout with TOOL_REVIEW_RESULT on its own line followed by exactly one JSON object with reviewId, createdAt, and decisions."
+}

package/node_modules/@llblab/pi-actors/dist/skills/swarm/SKILL.md

@@ -0,0 +1,84 @@
+---
+name: swarm
+description: Use when work needs multiple actors or subagents for independent implementation, artifact generation, review, delegated audit, research, or coordinated decomposition and integration.
+---
+
+# Swarm
+
+Use multi-actor execution only when at least two scopes or evidence lenses are meaningfully independent and parallelism, clean-context judgement, or quorum confidence is worth the coordination overhead. Do not swarm a task that one bounded agent can complete safely, a task whose architecture is still unsettled, or concurrent mutations of one shared contract.
+
+Read `actors` first for generic Recipe, spawn, Run, Trace, Control, artifact, and lifecycle operation. This Skill owns only multi-actor methodology: decomposition, scope ownership, independence, synthesis, integration, and completion proof.
+
+## Coordinator topology
+
+A swarm can be coordinated without an external gateway. In this model the current host agent is the declarative control plane, the actor kernel creates explicit participant Runs, and companion transports provide ingress or presence without owning hidden agent creation. The coordinator retains user authority, global context, decomposition, shared-surface ownership, integration, and final validation; participants own bounded concrete tasks and report evidence.
+
+This resembles gateway orchestration in dependency direction but not in ownership: the coordinator is itself an agent instance with inspectable Runs, not an infrastructure service that implicitly creates sessions. Preserve that distinction in prompts, docs, recovery, and target routing.
+
+Once work is delegated, keep the coordinator available for decisions and integration instead of duplicating participant implementation. Wait for terminal follow-up by default; use meaningful attention or evidence-based timers for overdue work rather than a tight inspection loop.
+
+## Reasoning allocation
+
+Allocate reasoning by role instead of making one long thread implement and judge itself:
+
+- Bounded implementation/authorship participants default to reasoning off when the task card fixes scope, invariants, checks, and escalation. Enable reasoning only when unresolved diagnosis or local design judgement is part of their assignment.
+- Reviewers default to independent medium reasoning and clean context. For consequential work, several reviewers with distinct lenses or repeated independent judgement usually provide better error discovery than increasing one author's reasoning and relying on self-review.
+- Synthesizers and integrators use medium reasoning because they reconcile evidence, conflicts, shared contracts, and retained state.
+- The coordinator decides whether review fanout is worth its cost, preserves dissent, and never treats reviewer count as evidence quality by itself.
+
+Do not change a running participant's profile merely because policy changed. Replace or add a later independent review only when fresh evidence is still needed.
+
+## Choose the shape
+
+| Need | Shape | Primary Recipe |
+| --- | --- | --- |
+| Different risk lenses on one target | Lens swarm | `swarm/lens-review` |
+| Independent judges for one exact claim | Quorum | `swarm/quorum-review` |
+| Evidence map plus contradiction-preserving synthesis | Research swarm | `swarm/research-synthesis` |
+| Competing architecture directions and one smallest next slice | Architecture swarm | `swarm/architect` |
+| Bounded implementation assignment with scope critique | Development tasking | `swarm/development-tasking` |
+| Multi-lens ship/readiness verdict | Readiness review | `swarm/review-readiness` |
+
+Use different lenses for breadth and repeated independent judges for confidence. Combine both only for high-stakes work where the added cost is justified. `swarm/subagent-*` Recipes are maintained composition components; start from a primary Recipe unless building an intentional custom composition.
+
+## Coordinator protocol
+
+The coordinator owns the whole result even when participants choose local implementation details.
+
+1. State the goal, non-goals, evidence standard, integration owner, and stop condition.
+2. Partition work into disjoint read or write scopes. Give shared contracts one owner.
+3. Give each participant a bounded task card with allowed scope, avoided scope, expected artifact, checks, and escalation rule.
+4. Assign each participant an explicit execution profile and isolation mode under the reasoning-allocation contract.
+5. Preflight required model/tool access before expensive fanout.
+6. Launch independent work without cross-contaminating lenses. Do not let participants silently expand scope.
+7. Preserve every terminal result, including failures, disagreements, and partial evidence; avoid doing participant work in the coordinator while a valid owner remains active.
+8. Merge through one named synthesizer or integrator. Resolve conflicts from explicit intent and invariants, not textual convenience.
+9. Run fresh integrated validation and, for consequential outputs, an independent post-merge review.
+10. Report complete, degraded, or insufficient-data status honestly; name residual owners and next actions.
+
+## Scope and coordination rules
+
+- One writable scope has one owner. Parallel readers may share a stable target.
+- Public contracts, schemas, central configuration, and integration surfaces require exclusive ownership.
+- Concurrent writers use disjoint paths, isolated worktrees, or declared patch/artifact outputs.
+- Shared ledgers, lockfiles, generated contracts, metadata, schemas, release surfaces, and cross-domain configuration belong to one named integrator unless a task card transfers one surface to another exclusive owner.
+- Participants record shared-surface and other out-of-scope needs in handoff instead of editing them opportunistically.
+- Reasoning and model profiles are task-card inputs, not implicit properties of the whole swarm.
+- Coordinator checkpoints are bounded decision requests, not free-form actor chat.
+- Locks support scope ownership but do not replace coordinator judgement. Every lock must be bounded and releasable.
+- One integrator owns merge order, conflict resolution, and final validation.
+- A zero-conflict merge is not proof of semantic compatibility.
+
+## Evidence and quorum rules
+
+- Every material finding traces to inspected evidence or explicit uncertainty.
+- Preserve minority high-impact findings and contradictions; consensus does not erase them.
+- Keep reviewer evidence separate from merger findings.
+- If successful evidence is below the requested threshold, return degraded or insufficient data instead of inventing quorum.
+- Use a clean-context merger for serious quorum work. Use a fresh post-merge reviewer when the result drives code, security, architecture, money, governance, migrations, or release decisions.
+
+See [review swarms](./references/review-swarms.md) for lens, quorum, synthesis, and conflict-evidence detail. See [development swarms](./references/development-swarm.md) for task cards, write ownership, handoffs, conflict reports, and integration.
+
+## Stop rules
+
+Stop or replan when scopes overlap, a participant needs an undeclared shared contract, evidence cannot meet the threshold, provider/tool preflight fails, conflict changes the architecture, no integrator owns the result, or integrated validation is unavailable. Do not compensate with extra agents, repeated blind retries, shared mutable work, or coordinator-written consensus unsupported by participant evidence.

package/node_modules/@llblab/pi-actors/dist/skills/swarm/recipes/architect.json

@@ -0,0 +1,82 @@
+{
+  "async": true,
+  "imports": {
+    "reviewer": "swarm/subagent-review",
+    "critic": "swarm/subagent-critic",
+    "verifier": "swarm/subagent-verify",
+    "merger": "swarm/subagent-merge",
+    "normalizer": "swarm/subagent-normalize"
+  },
+  "args": [
+    "problem:string",
+    "lenses:array",
+    "constraints:string",
+    "model:string",
+    "tools:string"
+  ],
+  "defaults": {
+    "lenses": [
+      "architecture",
+      "implementation pragmatism",
+      "operator UX",
+      "failure modes"
+    ],
+    "constraints": "Prefer small composable pieces, explicit contracts, and no broad workflow DSL.",
+    "tools": ""
+  },
+  "template": [
+    {
+      "parallel": true,
+      "repeat": "{lenses.length}",
+      "failure": "branch",
+      "template": {
+        "name": "reviewer",
+        "values": {
+          "scope": "{problem}",
+          "lens": "{lenses[index]}",
+          "instructions": "Propose one architecture direction and one smallest next slice within constraints: {constraints}",
+          "model": "{model}",
+          "thinking": "medium",
+          "tools": "{tools}"
+        }
+      }
+    },
+    {
+      "name": "critic",
+      "values": {
+        "target": "Use proposed directions from stdin for problem: {problem}",
+        "constraints": "{constraints}",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    },
+    {
+      "name": "verifier",
+      "values": {
+        "claim": "The proposed architecture can be advanced by a small validated next slice.",
+        "evidence": "Use proposed directions and critique from stdin.",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    },
+    {
+      "name": "merger",
+      "values": {
+        "inputs": "Use brainstorm, critique, and verification outputs from stdin.",
+        "mode": "risk-first",
+        "risk_policy": "Prefer a feasible smallest next slice over ambitious completeness.",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    },
+    {
+      "name": "normalizer",
+      "values": {
+        "input": "Use architecture synthesis from stdin.",
+        "format": "Markdown sections: Direction, Rejected Alternatives, Risks, Component Contracts, Next Slice, Validation Gates.",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    }
+  ]
+}

package/node_modules/@llblab/pi-actors/dist/skills/swarm/recipes/checkpoint-continuation.json

@@ -0,0 +1,51 @@
+{
+  "async": true,
+  "imports": {
+    "checkpoint": "swarm/subagent-checkpoint",
+    "followup": "swarm/subagent-followup",
+    "normalizer": "swarm/subagent-normalize"
+  },
+  "args": [
+    "state:string",
+    "question:string",
+    "reply:string",
+    "mode:enum(same-context,degraded-new-branch)",
+    "model:string",
+    "tools:string"
+  ],
+  "defaults": {
+    "reply": "No coordinator reply provided yet; produce a checkpoint artifact only.",
+    "mode": "degraded-new-branch",
+    "tools": ""
+  },
+  "template": [
+    {
+      "name": "checkpoint",
+      "values": {
+        "state": "{state}",
+        "question": "{question}",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    },
+    {
+      "name": "followup",
+      "values": {
+        "checkpoint": "Use checkpoint output from stdin.",
+        "reply": "{reply}",
+        "mode": "{mode}",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    },
+    {
+      "name": "normalizer",
+      "values": {
+        "input": "Use checkpoint and follow-up outputs from stdin.",
+        "format": "Markdown sections: Checkpoint, Coordinator Reply, Continuation Mode, Lost Context, Next Actions.",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    }
+  ]
+}

package/node_modules/@llblab/pi-actors/dist/skills/swarm/recipes/development-tasking.json

@@ -0,0 +1,67 @@
+{
+  "async": true,
+  "imports": {
+    "planner": "swarm/subagent-plan",
+    "task_card": "swarm/subagent-task-card",
+    "critic": "swarm/subagent-critic",
+    "normalizer": "swarm/subagent-normalize"
+  },
+  "args": [
+    "goal:string",
+    "allowed:string",
+    "avoided:string",
+    "checks:string",
+    "constraints:string",
+    "model:string",
+    "tools:string"
+  ],
+  "defaults": {
+    "allowed": "Caller must provide allowed files, directories, or logical scopes before implementation.",
+    "avoided": "Avoid unrelated files, public contracts, generated artifacts, and broad refactors unless explicitly allowed.",
+    "checks": "Smallest validation that proves the task card is complete.",
+    "constraints": "Split into bounded mutation zones and preserve integrator ownership of shared contracts.",
+    "tools": ""
+  },
+  "template": [
+    {
+      "name": "planner",
+      "values": {
+        "goal": "{goal}",
+        "scope": "Allowed: {allowed}. Avoided: {avoided}.",
+        "constraints": "{constraints}",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    },
+    {
+      "name": "task_card",
+      "values": {
+        "goal": "Use plan from stdin to create one bounded implementation assignment for: {goal}",
+        "allowed": "{allowed}",
+        "avoided": "{avoided}",
+        "checks": "{checks}",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    },
+    {
+      "name": "critic",
+      "values": {
+        "target": "Use generated plan and task card from stdin.",
+        "constraints": "{constraints}",
+        "risk_policy": "Find unsafe scope expansion, missing ownership, and weak validation gates.",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    },
+    {
+      "name": "normalizer",
+      "values": {
+        "input": "Use plan, task card, and critique from stdin.",
+        "format": "Markdown sections: Task Card, Allowed Files, Avoid Files, Checks, Risks, Integrator Notes.",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    }
+  ]
+}

package/node_modules/@llblab/pi-actors/dist/skills/swarm/recipes/lens-review.json

@@ -0,0 +1,63 @@
+{
+  "description": "General-purpose multi-lens review swarm. Launches independent reviewers by lens, then verifies, merges, judges, and normalizes the result.",
+  "async": true,
+  "imports": {
+    "coordinator": "swarm/subagent-review-coordinator"
+  },
+  "args": [
+    "scope:string",
+    "lenses:array",
+    "model:string",
+    "thinking:string",
+    "tools:string",
+    "subagent_ttl_ms:int",
+    "reviewer_concurrency:string",
+    "min_successful_reviewers:int",
+    "merge_policy:string",
+    "claim:string",
+    "evidence_policy:string",
+    "risk_policy:string",
+    "output_format:string"
+  ],
+  "defaults": {
+    "lenses": [
+      "correctness",
+      "architecture",
+      "operator UX"
+    ],
+    "model": "{current_model}",
+    "thinking": "{current_thinking}",
+    "tools": "",
+    "subagent_ttl_ms": 600000,
+    "reviewer_concurrency": "",
+    "min_successful_reviewers": 1,
+    "merge_policy": "Preserve partial reviewer reports, branch status, and degraded confidence markers. If reviewer evidence is below threshold, do not invent consensus.",
+    "claim": "The reviewed scope has clear, evidence-based findings, risks, and recommended next actions.",
+    "evidence_policy": "Cite inspected files, command outputs, or explicit uncertainty for every material claim.",
+    "risk_policy": "Preserve minority high-impact risks and ideas; separate confirmed issues from hypotheses.",
+    "output_format": "Markdown sections: Status (complete/degraded/insufficient_data), Summary, Consensus Findings, Minority Findings, Verification, Judge Notes, Risks, Next Actions."
+  },
+  "template": [
+    {
+      "name": "coordinator",
+      "values": {
+        "scope": "{scope}",
+        "lenses": "{lenses}",
+        "claim": "{claim}",
+        "reviewer_model": "{model}",
+        "verifier_model": "{model}",
+        "merger_model": "{model}",
+        "judge_model": "{model}",
+        "thinking": "{thinking}",
+        "tools": "{tools}",
+        "subagent_ttl_ms": "{subagent_ttl_ms}",
+        "reviewer_concurrency": "{reviewer_concurrency}",
+        "min_successful_reviewers": "{min_successful_reviewers}",
+        "merge_policy": "{merge_policy}",
+        "evidence_policy": "{evidence_policy}",
+        "risk_policy": "{risk_policy}",
+        "output_format": "{output_format}"
+      }
+    }
+  ]
+}

package/node_modules/@llblab/pi-actors/dist/skills/swarm/recipes/quorum-review.json

@@ -0,0 +1,65 @@
+{
+  "async": true,
+  "imports": {
+    "subagent": "swarm/subagent-prompt",
+    "merger": "swarm/subagent-merge",
+    "judge": "swarm/subagent-judge",
+    "normalizer": "swarm/subagent-normalize"
+  },
+  "args": [
+    "prompt:string",
+    "models:array",
+    "merge_mode:enum(consensus-first,risk-first)",
+    "merger_model:string",
+    "judge_model:string",
+    "tools:string"
+  ],
+  "defaults": {
+    "merge_mode": "consensus-first",
+    "tools": ""
+  },
+  "template": [
+    {
+      "parallel": true,
+      "repeat": "{models.length}",
+      "failure": "branch",
+      "template": {
+        "name": "subagent",
+        "values": {
+          "prompt": "{prompt}",
+          "model": "{models[index]}",
+          "tools": "{tools}",
+          "output_format": "Vote-style Markdown: Verdict, Rationale, Evidence, Confidence."
+        }
+      }
+    },
+    {
+      "name": "merger",
+      "values": {
+        "inputs": "Use quorum outputs from stdin.",
+        "mode": "{merge_mode}",
+        "model": "{merger_model}",
+        "tools": "{tools}",
+        "risk_policy": "Preserve vote shape, minority severe findings, and confidence disagreements."
+      }
+    },
+    {
+      "name": "judge",
+      "values": {
+        "report": "Use merged quorum report from stdin.",
+        "evidence": "Use raw quorum outputs and merged report from stdin.",
+        "model": "{judge_model}",
+        "tools": "{tools}"
+      }
+    },
+    {
+      "name": "normalizer",
+      "values": {
+        "input": "Use merged and judged quorum output from stdin.",
+        "format": "Markdown sections: Verdict, Vote Shape, Consensus, Minority Findings, Confidence, Required Action.",
+        "model": "{merger_model}",
+        "tools": "{tools}"
+      }
+    }
+  ]
+}

package/node_modules/@llblab/pi-actors/dist/skills/swarm/recipes/research-synthesis.json

@@ -0,0 +1,81 @@
+{
+  "async": true,
+  "imports": {
+    "planner": "swarm/subagent-plan",
+    "evidence": "swarm/subagent-evidence-map",
+    "contradictions": "swarm/subagent-contradiction-map",
+    "verifier": "swarm/subagent-verify",
+    "merger": "swarm/subagent-merge",
+    "normalizer": "swarm/subagent-normalize"
+  },
+  "args": [
+    "question:string",
+    "sources:string",
+    "constraints:string",
+    "model:string",
+    "tools:string"
+  ],
+  "defaults": {
+    "sources": "Use provided source notes or artifact paths only; mark missing evidence.",
+    "constraints": "Bound the question, preserve contradictions, and avoid publication-style certainty when evidence is weak.",
+    "tools": ""
+  },
+  "template": [
+    {
+      "name": "planner",
+      "values": {
+        "goal": "Answer research question: {question}",
+        "scope": "Sources: {sources}",
+        "constraints": "{constraints}",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    },
+    {
+      "name": "evidence",
+      "values": {
+        "question": "{question}",
+        "sources": "{sources}",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    },
+    {
+      "name": "contradictions",
+      "values": {
+        "inputs": "Use plan and evidence map from stdin.",
+        "focus": "Contradictions, weak evidence, and missing source classes for question: {question}",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    },
+    {
+      "name": "verifier",
+      "values": {
+        "claim": "The synthesis can answer the research question within the stated constraints.",
+        "evidence": "Use plan, evidence map, and contradiction map from stdin.",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    },
+    {
+      "name": "merger",
+      "values": {
+        "inputs": "Use all previous research outputs from stdin.",
+        "mode": "risk-first",
+        "risk_policy": "Preserve uncertainty and contradictions over forced consensus.",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    },
+    {
+      "name": "normalizer",
+      "values": {
+        "input": "Use merged research synthesis from stdin.",
+        "format": "Markdown sections: Question, Evidence Map, Synthesis, Contradictions, Confidence, Missing Evidence, Next Evidence Slice.",
+        "model": "{model}",
+        "tools": "{tools}"
+      }
+    }
+  ]
+}

package/node_modules/@llblab/pi-actors/dist/skills/swarm/recipes/review-readiness.json

@@ -0,0 +1,59 @@
+{
+  "async": true,
+  "imports": {
+    "coordinator": "swarm/subagent-review-coordinator"
+  },
+  "args": [
+    "scope:string",
+    "lenses:array",
+    "release_gate:string",
+    "reviewer_model:string",
+    "verifier_model:string",
+    "merger_model:string",
+    "judge_model:string",
+    "thinking:string",
+    "tools:string",
+    "subagent_ttl_ms:int",
+    "reviewer_concurrency:string",
+    "min_successful_reviewers:int",
+    "merge_policy:string"
+  ],
+  "defaults": {
+    "lenses": [
+      "correctness",
+      "tests",
+      "release",
+      "operator UX"
+    ],
+    "release_gate": "The scope is safe to ship or hand to the next release step.",
+    "reviewer_model": "{current_model}",
+    "verifier_model": "{current_model}",
+    "merger_model": "{current_model}",
+    "judge_model": "{current_model}",
+    "thinking": "{current_thinking}",
+    "tools": "",
+    "subagent_ttl_ms": 600000,
+    "reviewer_concurrency": "",
+    "min_successful_reviewers": 1,
+    "merge_policy": "Preserve partial reviewer reports, branch status, and degraded confidence markers. If reviewer evidence is below threshold, do not invent consensus."
+  },
+  "template": {
+    "name": "coordinator",
+    "values": {
+      "scope": "{scope}",
+      "lenses": "{lenses}",
+      "claim": "{release_gate}",
+      "reviewer_model": "{reviewer_model}",
+      "verifier_model": "{verifier_model}",
+      "merger_model": "{merger_model}",
+      "judge_model": "{judge_model}",
+      "thinking": "{thinking}",
+      "tools": "{tools}",
+      "subagent_ttl_ms": "{subagent_ttl_ms}",
+      "reviewer_concurrency": "{reviewer_concurrency}",
+      "min_successful_reviewers": "{min_successful_reviewers}",
+      "merge_policy": "{merge_policy}",
+      "output_format": "Markdown sections: Status (complete/degraded/insufficient_data), Ship Verdict, Blocking Findings, Nonblocking Risks, Evidence, Required Follow-up."
+    }
+  }
+}

package/node_modules/@llblab/pi-actors/dist/skills/swarm/recipes/subagent-artifact.json

@@ -0,0 +1,21 @@
+{
+  "async": true,
+  "args": [
+    "input:string",
+    "artifact_path:path",
+    "artifact_format:string",
+    "write_policy:string",
+    "validation_policy:string",
+    "model:string",
+    "thinking:string",
+    "tools:string"
+  ],
+  "defaults": {
+    "artifact_format": "Markdown with stable headings and preserved evidence.",
+    "write_policy": "Prepare content for the artifact path and state whether the caller should write, append, or review before writing.",
+    "validation_policy": "List checks the caller should run before accepting the artifact; distinguish required checks from optional review.",
+    "thinking": "off",
+    "tools": ""
+  },
+  "template": "pi -p --model {model} --thinking {thinking} {tools?--tools:--no-tools} {tools} Prepare a durable artifact. Target path: {artifact_path}. Format: {artifact_format}. Write policy: {write_policy}. Validation policy: {validation_policy}. Input: {input}"
+}

package/node_modules/@llblab/pi-actors/dist/skills/swarm/recipes/subagent-checkpoint.json

@@ -0,0 +1,21 @@
+{
+  "async": true,
+  "args": [
+    "state:string",
+    "question:string",
+    "options:string",
+    "resume_policy:string",
+    "output_format:string",
+    "model:string",
+    "thinking:string",
+    "tools:string"
+  ],
+  "defaults": {
+    "options": "List options considered, recommendation, and risk if the coordinator does not answer.",
+    "resume_policy": "Do not claim same-context resume unless the adapter provides a resumable session id; otherwise emit degraded artifact state.",
+    "output_format": "Markdown sections: Preserved State, Question, Options, Recommendation, Risk If Guessed, Resume Mode.",
+    "thinking": "off",
+    "tools": ""
+  },
+  "template": "pi -p --model {model} --thinking {thinking} {tools?--tools:--no-tools} {tools} Create a bounded coordinator checkpoint. State to preserve: {state}. Question: {question}. Options/risk format: {options}. Resume policy: {resume_policy}. Output format: {output_format}"
+}

package/node_modules/@llblab/pi-actors/dist/skills/swarm/recipes/subagent-conflict-report.json

@@ -0,0 +1,19 @@
+{
+  "async": true,
+  "args": [
+    "change:string",
+    "conflict:string",
+    "preserve_policy:string",
+    "output_format:string",
+    "model:string",
+    "thinking:string",
+    "tools:string"
+  ],
+  "defaults": {
+    "preserve_policy": "Describe what must be preserved, what can be discarded, and suggested resolver action.",
+    "output_format": "Markdown sections: Conflict, My Changes, Other Changes, Preserve, Discard, Suggested Resolution, Risks.",
+    "thinking": "off",
+    "tools": ""
+  },
+  "template": "pi -p --model {model} --thinking {thinking} {tools?--tools:--no-tools} {tools} Write a bounded conflict report. My change: {change}. Conflict: {conflict}. Preserve policy: {preserve_policy}. Output format: {output_format}"
+}

package/node_modules/@llblab/pi-actors/dist/skills/swarm/recipes/subagent-contradiction-map.json

@@ -0,0 +1,20 @@
+{
+  "async": true,
+  "args": [
+    "inputs:string",
+    "focus:string",
+    "resolution_policy:string",
+    "output_format:string",
+    "model:string",
+    "thinking:string",
+    "tools:string"
+  ],
+  "defaults": {
+    "focus": "Find contradictions, unresolved assumptions, incompatible recommendations, and evidence gaps.",
+    "resolution_policy": "Do not force consensus; preserve unresolved conflicts and name what evidence would resolve them.",
+    "output_format": "Markdown sections: Contradictions, Affected Claims, Resolution Evidence, Unresolved Risks.",
+    "thinking": "medium",
+    "tools": ""
+  },
+  "template": "pi -p --model {model} --thinking {thinking} {tools?--tools:--no-tools} {tools} Map contradictions in these inputs: {inputs}. Focus: {focus}. Resolution policy: {resolution_policy}. Output format: {output_format}"
+}