Keeps this page in sync as the body changes. Pause it any time for a quieter view.
Path /specs/coherence-network-agent-pipeline
Last refresh never
... 4 more in Requirements section below
inventory_source local | spec_api /api/spec-registry/coherence-network-agent-pipeline | registry_updated 2026-04-09T03:10:08.851505Z
potential_value 0.00 | actual_value 1.00 | value_gap 0.00
estimated_cost 0.00 | actual_cost 1.00 | cost_gap 1.00
estimated_roi 0.00 | actual_roi 1.00
Missing contributor linkage. Submit a change request with contributor attribution.
Open process view for this spec
task_ids -
branches -
source_files specs/agent-orchestration-api.md, specs/attention-heuristics-pipeline-status.md, specs/coherence-network-agent-pipeline.md, specs/pipeline-observability-and-auto-review.md, specs/project-manager-pipeline.md, specs/split-review-deploy-verify-phases.md
evidence_refs -
Open implementation view for this spec
implementation_refs spec-registry:agent-orchestration-api, spec-registry:attention-heuristics-pipeline-status, spec-registry:coherence-network-agent-pipeline, spec-registry:pipeline-observability-and-auto-review, spec-registry:project-manager-pipeline, spec-registry:split-review-deploy-verify-phases
lineage_ids -
public_endpoints -
summary The project currently requires manual orchestration to move ideas through the development pipeline — an operator must identify which opportunities are highest-ROI, create tasks, monitor execution, and advance idea stages. This spec defines a persistent background loop ("agent pipeline") that continuously selects the highest-ROI pending task candidate from the portfolio, generates and executes work via `local_runner.py`, and auto-advances ideas through lifecycle stages (none → specced → implement
process_summary R1: A new `api/scripts/agent_pipeline.py` script implements a persistent background loop that polls the idea portfolio a; R2: Each poll cycle builds a pending-task candidate set from idea stages and ranks candidates by ROI score (coherence_sc; R3: For each selected idea, the pipeline determines the next required task type based on the idea's current stage: `none; R4: The pipeline creates an `AgentTask` via the agent API (`POST /api/agent/tasks`) with the idea's context, then delega; R5: On successful task completion, the pipeline calls the idea auto-advance endpoint (`POST /api/ideas/{idea_id}/advance
pseudocode_summary -
implementation_summary api/app/services/pipeline_service.py (pipeline orchestration); api/app/services/pipeline_pulse_service.py (compute_pulse()); api/app/services/pipeline_advance_service.py (auto-advance, auto-retry, escalation)