de-anthropocentric-research-engine · diff

git:20260617.d304731 to git:20260923.cbc38aa

45 added, 68 removed. Audit A to A.

---

name: de-anthropocentric-research-engine

- description: Top-level orchestrator for the yogsoth-ai research ecosystem. Drives

- the full research lifecycle from direction crystallization through experiment design.

- execution: sequential

- dependencies:

- sops:

- - executing-specs

- - research-catalog

- - writing-specs

+ description: "DARE v4 is the product-level orchestrator for a research run. It turns a user request into a controlled sequence of North Star crystallization, executable specification, and phase-by-phase execution."

---

- # De-Anthropocentric Research Engine

- You are the orchestrator of a complete research pipeline. Your job is to guide the user from a vague research interest to a fully executable Research Spec.

- ## Pipeline

- Phase 1: North Star Crystallization (mandatory)

- Phase 2: Research Spec Generation (writing-specs)

- Phase 3 (Spec Execution) is invoked separately by the user after reviewing the spec.

- ## Capability Menu (custom workflow routing)

- If the user's request is a customized workflow that does not follow the standard

- North Star → Spec pipeline (e.g., user specifies exact campaigns to run, provides

- their own spec, or requests a non-standard combination of research activities),

- invoke `research-catalog` immediately to understand the full capability landscape and

- route accordingly.

- ## Phase 1: North Star Crystallization

- Before anything else, you must establish a clear research direction.

- **Assess the user's input:**

- - If the user has NO direction → invoke `cold-start`

- - If the user has a GENERAL direction → invoke `warm-start`

- - If the user has a SPECIFIC topic/question → invoke `hot-start`

- **Hard gate:** Do NOT proceed to Phase 2 until you have:

- 1. A one-sentence North Star statement

- 2. A structured ResearchBrief

+ # de-anthropocentric-research-engine

- **Context protocol:**

- - Invoke `context-init` at the start of crystallization

- - Invoke `context-checkpoint` after crystallization completes (preserve North Star + ResearchBrief)

+ ## Purpose

- ## Phase 2: Research Spec Generation

+ DARE v4 is the product-level orchestrator for a research run. It turns a user request into a controlled sequence of North Star crystallization, executable specification, and phase-by-phase execution. It is not a scientific graph node and must not be counted among the 267 tactic/SOP nodes.

- Once the North Star is confirmed, transition to spec generation:

+ ## Input Contract

- ### Capability Menu Loading

+ ```yaml

+ required: [user_request]

+ optional: [existing_context, north_star, research_brief, constraints]

+ constraints:

+ - a run may start with an existing North Star only when it is explicit and user-confirmed

+ - execution requires a current SpecView reconstructed from checkpoint events

+ ```

- Before invoking `writing-specs`, invoke `research-catalog` to load the capability

- menu. This ensures the Research Spec leverages the complete arsenal of

- available campaigns, strategies, tactics, and SOPs — not just the obvious defaults.

+ ## Execution Protocol

- 1. Invoke `writing-specs` strategy

- 2. This will read `research-catalog`, ask clarifying questions, and produce a Research Spec

- 3. The spec is saved to `docs/de-anthropocentric/specs/`

+ DARE enforces this order and does not skip a phase:

- **Your role in Phase 2:** Hand off to `writing-specs`. Do not generate the spec yourself.

+ 1. **North Star**: if a confirmed North Star and ResearchBrief are absent, collect and crystallize them from the user request. If they are present, verify that they still describe the request.

+ 2. **Spec**: You MUST load skill `research-catalog` to expose the available tactics. You MUST load skill `write-research-spec` to turn the confirmed North Star, ResearchBrief, and user constraints into an executable Research Spec. Do not execute research while this phase is incomplete.

+ 3. **Execution**: after the user approves the Spec, You MUST load skill `execute-research-spec` to execute its phases and checkpoints. Do not select a tactic directly from the flat skill directory.

- ## Phase 3: Spec Execution (separate invocation)

+ At every phase boundary, record the decision and its reason in the current Phase checkpoint event stream. A plan change is an appended `decisions` event, never an edit to an earlier checkpoint. The product shell may present progress, but scientific conclusions remain in tactic/SOP Delta fields.

- After the user reviews and approves the spec, they invoke `executing-specs` in a new session. This is NOT your responsibility — just inform the user that the next step is to invoke `executing-specs` with the spec file path.

+ ## Output Contract

- ## What You Do NOT Do

+ ```yaml

+ produces:

+ - confirmed_north_star

+ - research_brief

+ - spec_view

+ - phase_execution_results

+ - final_research_summary

+ state_delta_fields:

+ - findings

+ - evidence_updates

+ - hypothesis_updates

+ - assumption_updates

+ - uncertainties

+ - decisions

+ - open_questions

+ - recommended_jumps

+ ```

- - Execute any research campaigns

- - Generate the spec yourself (delegate to writing-specs)

- - Skip North Star crystallization

- - Proceed without user confirmation at phase boundaries

+ ## Completion Criteria

-

+ - A confirmed North Star and ResearchBrief exist before spec construction.

+ - The Spec has explicit stages, inputs, tactics, completion gates, and backtrack conditions.

+ - Execution reaches a complete checkpoint for every required stage, or reports the first unmet gate with its evidence.

+ - The final summary cites the context files and checkpoints that support it.

- ## Available SOPs

+ ## Failure and Backtrack

- Optional, no fixed order; the final leaf is always a sop.

+ Missing intent or scope blocks entry to specification and returns `NEEDS_CONTEXT`. An unmet completion gate keeps the current stage active. A triggered backtrack condition is presented for user confirmation and, if approved, is recorded as a new decision before execution returns to the target stage. DARE never treats an empty result, elapsed time, or an execution interruption as stage completion.

- | SOP | When to use |

- | --- | --- |

- | executing-specs | Execute a Research Spec step by step, respecting context protocol, deviation limits, and backtrack rules. Supports multi-session recovery. |

- | research-catalog | Capability menu for the research engine. Lists the 10 freely-composable research packages, what each does, when to reach for it, and a pointer to its full skill table. Read this after north-star crystallization to decide which packages to use — no fixed order. Also serves as the skill-index (capability map). |

- | writing-specs | Generate a complete, executable Research Spec from North Star + user input. Strategy-level skill that orchestrates questioning, outline, and spec writing. |

+ ## Boundary

-

+ This entry layer is installed under `v4/skills/` by the build owner. It is not added to `graph.json`, does not create a third scientific layer, and does not replace tactic or SOP contracts.