Add research-enabled responder capture

notes/fresh-workspace-handoff.md

@@ -49,6 +49,9 @@ path is:
 - Responder packet schema: `schemas/responder-packet.schema.json`
 - Responder output schema: `schemas/responder-output.schema.json`
 - Research-enabled Sella packet: `examples/responder-packets/scene-02-sanctuary-intake.sella_ren.packet.research.v0.json`
+- Research-enabled Sella capture: `experiments/responder-packets/cold-wake-sanctuary-intake-sella-research-enabled-v0.capture.json`
+- Research-enabled Sella mutation: `experiments/responder-packets/cold-wake-sanctuary-intake-sella-research-enabled-v0.mutation.json`
+- Research-enabled aftermath state: `examples/agent-state.cold-wake-story-lab.after-sella-research-conditions.json`
 - Packet builder: `tools/build_responder_packet.py`
 - Packet/output validators: `tools/validate_responder_packets.py`, `tools/validate_responder_outputs.py`
 
@@ -71,14 +74,13 @@ current gold-data route.
 
 ## Current Next Action
 
-Run a sandboxed no-fork research-enabled responder against
-`examples/responder-packets/scene-02-sanctuary-intake.sella_ren.packet.research.v0.json`,
-allowing only the declared AetheriaLore scope. Capture exact visible input,
-consulted refs, research summary, raw output, review labels, leakage audit, and
-any coordinator intervention. Review whether the output uses institutional lore
-and latent pressure without prompt parroting, omniscience, or mission-critical
-trauma-dumping. Do not use archived local-model prototype runners for gold
-responder data.
+The first no-fork research-enabled Sella responder sample now exists and has a
+reviewed mutation receipt. Tighten the capture seam next: future research-enabled
+gold samples should preserve auditable research traces, not only
+responder-reported consulted refs. Then generate the next sandboxed
+responder/coordinator sample from the updated Sella aftermath state, preserving
+exact visible input, raw output, review labels, leakage audit, and mutation
+receipts.
 
 ## Warnings
 

notes/ghostlight-current-system-map.md

@@ -52,8 +52,15 @@ Aetheria lore/source context
   `examples/responder-packets/scene-02-sanctuary-intake.sella_ren.packet.research.v0.json`
 - It uses `responder_scoped_repository_search`, allows only declared
   AetheriaLore scope, and requires `consulted_refs` plus `research_summary`.
-- The next useful sample should be generated through a no-fork sandboxed
-  responder, not through archived local-model prototype runners.
+- First research-enabled capture:
+  `experiments/responder-packets/cold-wake-sanctuary-intake-sella-research-enabled-v0.capture.json`
+- First research-enabled mutation receipt:
+  `experiments/responder-packets/cold-wake-sanctuary-intake-sella-research-enabled-v0.mutation.json`
+- First research-enabled aftermath state:
+  `examples/agent-state.cold-wake-story-lab.after-sella-research-conditions.json`
+- The capture is useful draft data, but the current subagent seam preserves
+  responder-reported consulted refs rather than a machine-verifiable research
+  trace. Fix that before producing many more research-enabled gold samples.
 
 ## Archived Receipts
 

notes/ghostlight-implementation-plan.md

@@ -20,16 +20,17 @@ sandboxed training-data pipeline for branching scenes and state consequences.
    - Preserve exact responder-visible input, hidden-context audit, allowed actions, source excerpts, output contract, and lore-access mode.
    - Preserve raw output, parsed output, review labels, consulted refs, research summary, leakage audit, and coordinator interventions.
 
-3. Generate the next research-enabled responder sample.
-   - Use `examples/responder-packets/scene-02-sanctuary-intake.sella_ren.packet.research.v0.json`.
-   - Run a no-fork sandboxed responder with only the packet and declared lore scope.
-   - Require actual lore consultation before answering.
-   - Review for institutional grounding, latent character pressure, non-omniscience, prompt parroting, and mission-critical trauma-dumping.
-   - Do not use archived local-model prototype runners for gold responder data.
+3. Tighten research-enabled responder capture.
+   - The first no-fork research-enabled Sella sample and reviewed mutation receipt now exist.
+   - Preserve `consulted_refs` and `research_summary` inside both raw responder output and capture metadata.
+   - Add an auditable research trace before scaling this lane; responder-reported consulted refs are useful but not enough for durable gold data.
+   - Continue rejecting archived local-model prototype runners for gold responder data unless the task is explicitly model plumbing.
 
-4. Materialize reviewed consequences.
-   - Convert accepted responder output into a mutation receipt.
-   - Update only reviewed scene state, memories, relationship stance, perceived overlays, and unresolved hooks.
+4. Generate the next reviewed sample.
+   - Start from `examples/agent-state.cold-wake-story-lab.after-sella-research-conditions.json`.
+   - Build the next projected local context and responder packet from reviewed state, not from coordinator memory.
+   - Run the next no-fork sandboxed responder/coordinator step with exact visible input, raw output, leakage audit, and review labels.
+   - Convert accepted output into mutation receipts.
    - Keep fuzzy social changes manual and auditable until appraiser/classifier models exist.
 
 5. Generalize the loop.