open-research-protocol 0.4.16 → 0.4.18

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/cli/orp.py CHANGED
@@ -8940,7 +8940,7 @@ def _about_payload() -> dict[str, Any]:
8940
8940
  "Frontier control is a built-in ORP ability exposed through `orp frontier ...`, separating the exact live point, the exact active milestone, the near structured checklist, and the farther major-version stack.",
8941
8941
  "Agent modes are lightweight optional overlays for taste, perspective shifts, and fresh movement; `orp mode nudge sleek-minimal-progressive --json` gives agents a deterministic reminder they can call on when they want a deeper, wider, top-down, or rotated lens without changing ORP's core artifact boundaries.",
8942
8942
  "Project/session linking is a built-in ORP ability exposed through `orp link ...` and stored machine-locally under `.git/orp/link/`.",
8943
- "Hosted secret inventory remains the canonical source of truth, while the local macOS Keychain cache is inspectable and usable through `orp secrets list`, `orp secrets ensure`, `orp secrets keychain-list`, `orp secrets sync-keychain`, and `orp secrets resolve --local-first`.",
8943
+ "Secrets are easiest to understand as saved keys and tokens: humans usually run `orp secrets add ...` and paste the value at the prompt, agents usually pipe the value with `--value-stdin`, and local macOS Keychain caching plus hosted sync are optional layers on top.",
8944
8944
  "Machine runner identity, heartbeat, hosted sync, prompt-job execution, and lease control are built into ORP through `orp runner status`, `orp runner enable`, `orp runner disable`, `orp runner heartbeat`, `orp runner sync`, `orp runner work`, `orp runner cancel`, and `orp runner retry`.",
8945
8945
  "Repo governance is built into ORP through `orp init`, `orp status`, `orp branch start`, `orp checkpoint create`, `orp backup`, `orp ready`, `orp doctor`, and `orp cleanup`.",
8946
8946
  "Hosted workspace operations are built directly into ORP under `orp workspaces ...`, plus the linked auth/ideas/feature/world/checkpoint/agent surfaces.",
@@ -9046,11 +9046,11 @@ def _home_payload(repo_root: Path, config_arg: str) -> dict[str, Any]:
9046
9046
  "command": "orp workspace list",
9047
9047
  },
9048
9048
  {
9049
- "label": "Print exact copyable crash-recovery commands for the main workspace",
9049
+ "label": "Inspect saved paths and exact recovery commands for the main workspace",
9050
9050
  "command": "orp workspace tabs main",
9051
9051
  },
9052
9052
  {
9053
- "label": "Add a new provider key interactively when you need one",
9053
+ "label": "Save a new API key or token interactively when you need one",
9054
9054
  "command": 'orp secrets add --alias <alias> --label "<label>" --provider <provider>',
9055
9055
  },
9056
9056
  {
@@ -9076,10 +9076,6 @@ def _home_payload(repo_root: Path, config_arg: str) -> dict[str, Any]:
9076
9076
  "label": "Inspect the saved workspace ledger inventory",
9077
9077
  "command": "orp workspace list",
9078
9078
  },
9079
- {
9080
- "label": "Print exact copyable crash-recovery commands for the main workspace",
9081
- "command": "orp workspace tabs main",
9082
- },
9083
9079
  {
9084
9080
  "label": "Inspect the saved tabs in the main workspace ledger",
9085
9081
  "command": "orp workspace tabs main",
@@ -9113,11 +9109,11 @@ def _home_payload(repo_root: Path, config_arg: str) -> dict[str, Any]:
9113
9109
  "command": "orp workspaces list --json",
9114
9110
  },
9115
9111
  {
9116
- "label": "Inspect the global ORP secret inventory",
9112
+ "label": "Inspect saved keys and tokens already known to ORP",
9117
9113
  "command": "orp secrets list --json",
9118
9114
  },
9119
9115
  {
9120
- "label": "Reuse a saved provider key or prompt for it and save it for this project",
9116
+ "label": "Reuse a saved key or prompt for it and save it for this project",
9121
9117
  "command": "orp secrets ensure --alias <alias> --provider <provider> --current-project --json",
9122
9118
  },
9123
9119
  {
@@ -9408,11 +9404,11 @@ def _home_payload(repo_root: Path, config_arg: str) -> dict[str, Any]:
9408
9404
  },
9409
9405
  {
9410
9406
  "id": "secrets",
9411
- "description": "Saved API keys and tokens, with create-if-missing flows, optional local macOS Keychain caching, and project-scoped resolution.",
9407
+ "description": "Saved API keys and tokens, with an interactive human flow, a stdin agent flow, optional local macOS Keychain caching, and optional hosted sync.",
9412
9408
  "entrypoints": [
9413
9409
  "orp secrets list --json",
9414
9410
  "orp secrets show <alias-or-id> --json",
9415
- "orp secrets add --alias <alias> --provider <provider> --current-project",
9411
+ 'orp secrets add --alias <alias> --label "<label>" --provider <provider>',
9416
9412
  "orp secrets ensure --alias <alias> --provider <provider> --current-project --json",
9417
9413
  "orp secrets keychain-list --json",
9418
9414
  "orp secrets keychain-show <alias-or-id> --json",
@@ -9614,14 +9610,34 @@ def _render_home_screen(payload: dict[str, Any]) -> str:
9614
9610
  lines.append("")
9615
9611
  lines.append("Command Families")
9616
9612
  if isinstance(abilities, list) and abilities:
9617
- for row in abilities:
9613
+ ability_map = {
9614
+ str(row.get("id", "")).strip(): row
9615
+ for row in abilities
9616
+ if isinstance(row, dict) and str(row.get("id", "")).strip()
9617
+ }
9618
+ visible_ability_ids = [
9619
+ "workspace",
9620
+ "secrets",
9621
+ "governance",
9622
+ "frontier",
9623
+ "schedule",
9624
+ "modes",
9625
+ "hosted",
9626
+ "discover",
9627
+ ]
9628
+ shown = 0
9629
+ for ability_id in visible_ability_ids:
9630
+ row = ability_map.get(ability_id)
9618
9631
  if not isinstance(row, dict):
9619
9632
  continue
9620
- ability_id = str(row.get("id", "")).strip()
9621
9633
  desc = _truncate(str(row.get("description", "")).strip())
9622
9634
  lines.append(f" - {ability_id}")
9623
9635
  if desc:
9624
9636
  lines.append(f" {desc}")
9637
+ shown += 1
9638
+ remaining = max(len(ability_map) - shown, 0)
9639
+ if remaining:
9640
+ lines.append(f" - ... and {remaining} more in `orp about --json`")
9625
9641
 
9626
9642
  lines.append("")
9627
9643
  lines.append("Collaboration")
@@ -9652,7 +9668,7 @@ def _render_home_screen(payload: dict[str, Any]) -> str:
9652
9668
  lines.append("")
9653
9669
  lines.append("Quick Actions")
9654
9670
  if isinstance(quick_actions, list):
9655
- for row in quick_actions[:14]:
9671
+ for row in quick_actions[:10]:
9656
9672
  if not isinstance(row, dict):
9657
9673
  continue
9658
9674
  label = str(row.get("label", "")).strip()
@@ -9661,7 +9677,7 @@ def _render_home_screen(payload: dict[str, Any]) -> str:
9661
9677
  continue
9662
9678
  lines.append(f" - {label}")
9663
9679
  lines.append(f" {command}")
9664
- remaining = max(len(quick_actions) - 14, 0)
9680
+ remaining = max(len(quick_actions) - 10, 0)
9665
9681
  if remaining:
9666
9682
  lines.append(f" - ... and {remaining} more in `orp home --json`")
9667
9683
 
@@ -18544,10 +18560,31 @@ def build_parser() -> argparse.ArgumentParser:
18544
18560
  add_json_flag(s_youtube_inspect)
18545
18561
  s_youtube_inspect.set_defaults(func=cmd_youtube_inspect, json_output=False)
18546
18562
 
18547
- s_secrets = sub.add_parser("secrets", help="Hosted secret store and project binding operations")
18563
+ s_secrets = sub.add_parser(
18564
+ "secrets",
18565
+ help="Save and reuse API keys and tokens locally, with optional hosted sync",
18566
+ description=(
18567
+ "ORP secrets are easiest to understand as saved keys and tokens.\n\n"
18568
+ "Human flow:\n"
18569
+ " 1. Run `orp secrets add ...`\n"
18570
+ " 2. Paste the value when ORP prompts `Secret value:`\n"
18571
+ " 3. Later run `orp secrets list` or `orp secrets resolve ...`\n\n"
18572
+ "Agent flow:\n"
18573
+ " - Pipe the value with `--value-stdin` instead of typing it interactively.\n\n"
18574
+ "Local macOS Keychain caching and hosted sync are optional layers on top."
18575
+ ),
18576
+ epilog=(
18577
+ "Examples:\n"
18578
+ " orp secrets add --alias openai-primary --label \"OpenAI Primary\" --provider openai\n"
18579
+ " printf '%s' 'sk-...' | orp secrets add --alias openai-primary --label \"OpenAI Primary\" --provider openai --value-stdin\n"
18580
+ " orp secrets list\n"
18581
+ " orp secrets resolve openai-primary --reveal"
18582
+ ),
18583
+ formatter_class=argparse.RawTextHelpFormatter,
18584
+ )
18548
18585
  secrets_sub = s_secrets.add_subparsers(dest="secrets_cmd", required=True)
18549
18586
 
18550
- s_secrets_list = secrets_sub.add_parser("list", help="List hosted secrets for the current user")
18587
+ s_secrets_list = secrets_sub.add_parser("list", help="List saved secrets known to ORP")
18551
18588
  s_secrets_list.add_argument("--provider", default="", help="Optional provider filter")
18552
18589
  add_secret_scope_flags(s_secrets_list)
18553
18590
  s_secrets_list.add_argument(
@@ -18559,13 +18596,16 @@ def build_parser() -> argparse.ArgumentParser:
18559
18596
  add_json_flag(s_secrets_list)
18560
18597
  s_secrets_list.set_defaults(func=cmd_secrets_list, json_output=False)
18561
18598
 
18562
- s_secrets_show = secrets_sub.add_parser("show", help="Show one hosted secret by alias or id")
18599
+ s_secrets_show = secrets_sub.add_parser("show", help="Show one saved secret by alias or id")
18563
18600
  s_secrets_show.add_argument("secret_ref", help="Secret alias or id")
18564
18601
  add_base_url_flag(s_secrets_show)
18565
18602
  add_json_flag(s_secrets_show)
18566
18603
  s_secrets_show.set_defaults(func=cmd_secrets_show, json_output=False)
18567
18604
 
18568
- s_secrets_add = secrets_sub.add_parser("add", help="Create a hosted secret")
18605
+ s_secrets_add = secrets_sub.add_parser(
18606
+ "add",
18607
+ help="Save a new secret; ORP prompts for the value unless you pass --value-stdin",
18608
+ )
18569
18609
  s_secrets_add.add_argument("--alias", required=True, help="Stable secret alias")
18570
18610
  s_secrets_add.add_argument("--label", required=True, help="Human label for the secret")
18571
18611
  s_secrets_add.add_argument("--provider", required=True, help="Provider slug, for example openai")
@@ -18596,7 +18636,7 @@ def build_parser() -> argparse.ArgumentParser:
18596
18636
 
18597
18637
  s_secrets_ensure = secrets_sub.add_parser(
18598
18638
  "ensure",
18599
- help="Use an existing hosted secret by alias or create it and bind it when missing",
18639
+ help="Reuse a saved secret or prompt for it and save it when missing",
18600
18640
  )
18601
18641
  s_secrets_ensure.add_argument("--alias", required=True, help="Stable secret alias")
18602
18642
  s_secrets_ensure.add_argument("--label", default="", help="Human label for create-if-missing flows")
@@ -18637,7 +18677,7 @@ def build_parser() -> argparse.ArgumentParser:
18637
18677
 
18638
18678
  s_secrets_keychain_list = secrets_sub.add_parser(
18639
18679
  "keychain-list",
18640
- help="List locally cached macOS Keychain secrets known to ORP on this machine",
18680
+ help="List local macOS Keychain copies known to ORP on this machine",
18641
18681
  )
18642
18682
  s_secrets_keychain_list.add_argument("--provider", default="", help="Optional provider filter")
18643
18683
  add_secret_scope_flags(s_secrets_keychain_list)
@@ -18646,7 +18686,7 @@ def build_parser() -> argparse.ArgumentParser:
18646
18686
 
18647
18687
  s_secrets_keychain_show = secrets_sub.add_parser(
18648
18688
  "keychain-show",
18649
- help="Show one locally cached macOS Keychain secret by alias or id",
18689
+ help="Show one local macOS Keychain copy by alias or id",
18650
18690
  )
18651
18691
  s_secrets_keychain_show.add_argument("secret_ref", help="Secret alias or id")
18652
18692
  s_secrets_keychain_show.add_argument(
@@ -18659,7 +18699,7 @@ def build_parser() -> argparse.ArgumentParser:
18659
18699
 
18660
18700
  s_secrets_sync_keychain = secrets_sub.add_parser(
18661
18701
  "sync-keychain",
18662
- help="Sync hosted secrets into the local macOS Keychain",
18702
+ help="Copy one saved secret into the local macOS Keychain",
18663
18703
  )
18664
18704
  s_secrets_sync_keychain.add_argument("secret_ref", nargs="?", default="", help="Optional secret alias or id")
18665
18705
  s_secrets_sync_keychain.add_argument("--provider", default="", help="Provider slug for project-scoped sync")
@@ -18673,7 +18713,7 @@ def build_parser() -> argparse.ArgumentParser:
18673
18713
  add_json_flag(s_secrets_sync_keychain)
18674
18714
  s_secrets_sync_keychain.set_defaults(func=cmd_secrets_sync_keychain, json_output=False)
18675
18715
 
18676
- s_secrets_update = secrets_sub.add_parser("update", help="Update one hosted secret")
18716
+ s_secrets_update = secrets_sub.add_parser("update", help="Update one saved secret")
18677
18717
  s_secrets_update.add_argument("secret_ref", help="Secret alias or id")
18678
18718
  s_secrets_update.add_argument("--alias", default=None, help="New alias")
18679
18719
  s_secrets_update.add_argument("--label", default=None, help="New label")
@@ -18702,13 +18742,13 @@ def build_parser() -> argparse.ArgumentParser:
18702
18742
  add_json_flag(s_secrets_update)
18703
18743
  s_secrets_update.set_defaults(func=cmd_secrets_update, json_output=False)
18704
18744
 
18705
- s_secrets_archive = secrets_sub.add_parser("archive", help="Archive one hosted secret")
18745
+ s_secrets_archive = secrets_sub.add_parser("archive", help="Archive one saved secret")
18706
18746
  s_secrets_archive.add_argument("secret_ref", help="Secret alias or id")
18707
18747
  add_base_url_flag(s_secrets_archive)
18708
18748
  add_json_flag(s_secrets_archive)
18709
18749
  s_secrets_archive.set_defaults(func=cmd_secrets_archive, json_output=False)
18710
18750
 
18711
- s_secrets_bind = secrets_sub.add_parser("bind", help="Bind one secret to a hosted project/world")
18751
+ s_secrets_bind = secrets_sub.add_parser("bind", help="Bind one saved secret to a hosted project/world")
18712
18752
  s_secrets_bind.add_argument("secret_ref", help="Secret alias or id")
18713
18753
  add_secret_scope_flags(s_secrets_bind)
18714
18754
  s_secrets_bind.add_argument("--purpose", default="", help="Optional project usage note")
@@ -18729,7 +18769,7 @@ def build_parser() -> argparse.ArgumentParser:
18729
18769
 
18730
18770
  s_secrets_resolve = secrets_sub.add_parser(
18731
18771
  "resolve",
18732
- help="Resolve one hosted secret by alias/id or by provider plus project scope",
18772
+ help="Resolve one saved secret by alias/id or by provider plus project scope",
18733
18773
  )
18734
18774
  s_secrets_resolve.add_argument("secret_ref", nargs="?", default="", help="Optional secret alias or id")
18735
18775
  s_secrets_resolve.add_argument("--provider", default="", help="Provider slug for project-scoped resolution")
@@ -243,10 +243,10 @@ Behavior:
243
243
 
244
244
  Formal schemas live in:
245
245
 
246
- - [link-project.schema.json](/Users/codymitchell/Documents/code/orp/spec/v1/link-project.schema.json)
247
- - [link-session.schema.json](/Users/codymitchell/Documents/code/orp/spec/v1/link-session.schema.json)
248
- - [runner-machine.schema.json](/Users/codymitchell/Documents/code/orp/spec/v1/runner-machine.schema.json)
249
- - [runner-runtime.schema.json](/Users/codymitchell/Documents/code/orp/spec/v1/runner-runtime.schema.json)
246
+ - [link-project.schema.json](../spec/v1/link-project.schema.json)
247
+ - [link-session.schema.json](../spec/v1/link-session.schema.json)
248
+ - [runner-machine.schema.json](../spec/v1/runner-machine.schema.json)
249
+ - [runner-runtime.schema.json](../spec/v1/runner-runtime.schema.json)
250
250
 
251
251
  Planned file locations and schema usage:
252
252
 
@@ -454,14 +454,14 @@ The CLI and Rust app now share one client-side project/session/runner contract,
454
454
  - [x] Add route/helper logging for failed poll/start/complete flows, lease mismatches, missing-routeable-session failures, and repeated retry patterns.
455
455
  - [x] Surface runner health in the Rust desktop app so operators can see online/syncing/working/error states locally.
456
456
  - [x] Add an internal rollout and recovery runbook:
457
- - [RUNNER_INTERNAL_OPERATIONS.md](/Users/codymitchell/Documents/code/orp/docs/RUNNER_INTERNAL_OPERATIONS.md)
457
+ - [RUNNER_INTERNAL_OPERATIONS.md](./RUNNER_INTERNAL_OPERATIONS.md)
458
458
  - [x] Deploy the hosted runner backend changes to the real internal environment.
459
459
  - [x] Run a live internal smoke on deployed infrastructure.
460
460
  - Completed on March 16, 2026 against `https://orp.earth`.
461
461
  - Verified `orp link project bind`, `orp link session register`, `orp runner enable`, `orp runner sync`, `orp checkpoint queue`, `orp runner work --once`, and `orp agent work --once`.
462
462
  - Confirmed the production `orp` checkpoint job `78cd459a-fc0b-451b-af06-be2d27379169` completed successfully and produced checkpoint response `41087b8b-9556-4ec1-90c6-eefb69bac585`.
463
463
  - [x] Add and verify a reusable Rust-side smoke harness for the desktop wrapper path.
464
- - Implemented at [runner_smoke.rs](/Users/codymitchell/Documents/code/orp-rust/src/bin/runner_smoke.rs).
464
+ - Implemented in the companion Rust workspace at `orp-rust/src/bin/runner_smoke.rs`.
465
465
  - Verified on March 16, 2026 against `https://orp.earth`.
466
466
  - Confirmed Rust-side smoke job `853a55f9-b0e5-42f7-8f2f-cdc8db1a354c` completed successfully and produced checkpoint response `6b5aee77-f176-4249-a127-978b987da946`.
467
467
 
@@ -51,7 +51,7 @@ Use this checklist when releasing ORP as the unified public CLI and product surf
51
51
  - `orp agent work --once --json` remains available as the compatibility path
52
52
  - Confirm the checkpoint response lands back in the hosted workspace.
53
53
  - Confirm the hosted operator console reflects the same lifecycle at `/dashboard/admin/runners`.
54
- - Use [RUNNER_INTERNAL_OPERATIONS.md](/Users/codymitchell/Documents/code/orp/docs/RUNNER_INTERNAL_OPERATIONS.md) for the internal rollout and recovery flow.
54
+ - Use [RUNNER_INTERNAL_OPERATIONS.md](./RUNNER_INTERNAL_OPERATIONS.md) for the internal rollout and recovery flow.
55
55
 
56
56
  ## 4. Package release
57
57
 
@@ -89,4 +89,4 @@ Use this checklist when releasing ORP as the unified public CLI and product surf
89
89
  - Keep the web app and CLI rollout loosely coupled.
90
90
  - Launch the ORP CLI first if the web app/domain transition is still in progress.
91
91
  - Do not change domain, auth, runner, and package names all in one step unless all staging checks are green.
92
- - Follow [ORP_WEB_DOMAIN_TRANSITION_PLAN.md](/Users/codymitchell/Documents/code/orp/docs/ORP_WEB_DOMAIN_TRANSITION_PLAN.md) for the hosted cutover sequence.
92
+ - Follow [ORP_WEB_DOMAIN_TRANSITION_PLAN.md](./ORP_WEB_DOMAIN_TRANSITION_PLAN.md) for the hosted cutover sequence.
@@ -5,12 +5,12 @@ the ORP Reasoning Kernel.
5
5
 
6
6
  Supporting artifact:
7
7
 
8
- - [docs/benchmarks/orp_reasoning_kernel_agent_pilot_v0_1.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_agent_pilot_v0_1.json)
8
+ - [docs/benchmarks/orp_reasoning_kernel_agent_pilot_v0_1.json](./benchmarks/orp_reasoning_kernel_agent_pilot_v0_1.json)
9
9
 
10
10
  Supporting corpus and harness:
11
11
 
12
- - [examples/kernel/comparison/comparison-corpus.json](/Volumes/Code_2TB/code/orp/examples/kernel/comparison/comparison-corpus.json)
13
- - [scripts/orp-kernel-agent-pilot.py](/Volumes/Code_2TB/code/orp/scripts/orp-kernel-agent-pilot.py)
12
+ - [examples/kernel/comparison/comparison-corpus.json](../examples/kernel/comparison/comparison-corpus.json)
13
+ - [scripts/orp-kernel-agent-pilot.py](../scripts/orp-kernel-agent-pilot.py)
14
14
 
15
15
  ## What This Pilot Measures
16
16
 
@@ -5,11 +5,11 @@ pilot for the live ORP kernel agent evaluation.
5
5
 
6
6
  Supporting artifact:
7
7
 
8
- - [docs/benchmarks/orp_reasoning_kernel_agent_replication_v0_2.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_agent_replication_v0_2.json)
8
+ - [docs/benchmarks/orp_reasoning_kernel_agent_replication_v0_2.json](./benchmarks/orp_reasoning_kernel_agent_replication_v0_2.json)
9
9
 
10
10
  Supporting harness:
11
11
 
12
- - [scripts/orp-kernel-agent-replication.py](/Volumes/Code_2TB/code/orp/scripts/orp-kernel-agent-replication.py)
12
+ - [scripts/orp-kernel-agent-replication.py](../scripts/orp-kernel-agent-replication.py)
13
13
 
14
14
  The harness now supports:
15
15
 
@@ -5,11 +5,11 @@ for the ORP Reasoning Kernel.
5
5
 
6
6
  Supporting artifact:
7
7
 
8
- - [docs/benchmarks/orp_reasoning_kernel_canonical_continuation_v0_1.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_canonical_continuation_v0_1.json)
8
+ - [docs/benchmarks/orp_reasoning_kernel_canonical_continuation_v0_1.json](./benchmarks/orp_reasoning_kernel_canonical_continuation_v0_1.json)
9
9
 
10
10
  Supporting harness:
11
11
 
12
- - [scripts/orp-kernel-canonical-continuation.py](/Volumes/Code_2TB/code/orp/scripts/orp-kernel-canonical-continuation.py)
12
+ - [scripts/orp-kernel-canonical-continuation.py](../scripts/orp-kernel-canonical-continuation.py)
13
13
 
14
14
  ## What This Pilot Measures
15
15
 
@@ -9,12 +9,12 @@ artifact styles:
9
9
 
10
10
  Supporting artifact:
11
11
 
12
- - [docs/benchmarks/orp_reasoning_kernel_comparison_v0_1.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_comparison_v0_1.json)
12
+ - [docs/benchmarks/orp_reasoning_kernel_comparison_v0_1.json](./benchmarks/orp_reasoning_kernel_comparison_v0_1.json)
13
13
 
14
14
  Supporting corpus and harness:
15
15
 
16
- - [examples/kernel/comparison/comparison-corpus.json](/Volumes/Code_2TB/code/orp/examples/kernel/comparison/comparison-corpus.json)
17
- - [scripts/orp-kernel-comparison.py](/Volumes/Code_2TB/code/orp/scripts/orp-kernel-comparison.py)
16
+ - [examples/kernel/comparison/comparison-corpus.json](../examples/kernel/comparison/comparison-corpus.json)
17
+ - [scripts/orp-kernel-comparison.py](../scripts/orp-kernel-comparison.py)
18
18
 
19
19
  ## What This Pilot Measures
20
20
 
@@ -91,7 +91,7 @@ This pilot does **not** prove that the kernel:
91
91
  - is universally superior across all teams or domains
92
92
 
93
93
  Those still require the larger studies in
94
- [docs/ORP_REASONING_KERNEL_EVALUATION_PLAN.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_EVALUATION_PLAN.md).
94
+ [docs/ORP_REASONING_KERNEL_EVALUATION_PLAN.md](./ORP_REASONING_KERNEL_EVALUATION_PLAN.md).
95
95
 
96
96
  ## Why The Scoring Is Structured This Way
97
97
 
@@ -5,15 +5,15 @@ Reasoning Kernel.
5
5
 
6
6
  Supporting artifact:
7
7
 
8
- - [docs/benchmarks/orp_reasoning_kernel_continuation_v0_1.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_continuation_v0_1.json)
8
+ - [docs/benchmarks/orp_reasoning_kernel_continuation_v0_1.json](./benchmarks/orp_reasoning_kernel_continuation_v0_1.json)
9
9
 
10
10
  Supporting harness:
11
11
 
12
- - [scripts/orp-kernel-continuation-pilot.py](/Volumes/Code_2TB/code/orp/scripts/orp-kernel-continuation-pilot.py)
12
+ - [scripts/orp-kernel-continuation-pilot.py](../scripts/orp-kernel-continuation-pilot.py)
13
13
 
14
14
  Related harder benchmark:
15
15
 
16
- - [docs/ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md)
16
+ - [docs/ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md](./ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md)
17
17
 
18
18
  ## What This Pilot Measures
19
19
 
@@ -13,14 +13,14 @@ to:
13
13
 
14
14
  Supporting references:
15
15
 
16
- - [docs/ORP_REASONING_KERNEL_COMPARISON_PILOT.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_COMPARISON_PILOT.md)
17
- - [docs/ORP_REASONING_KERNEL_PICKUP_PILOT.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_PICKUP_PILOT.md)
18
- - [docs/ORP_REASONING_KERNEL_AGENT_PILOT.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_AGENT_PILOT.md)
19
- - [docs/ORP_REASONING_KERNEL_AGENT_REPLICATION.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_AGENT_REPLICATION.md)
20
- - [docs/ORP_REASONING_KERNEL_CONTINUATION_PILOT.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_CONTINUATION_PILOT.md)
21
- - [docs/ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md)
22
- - [docs/ORP_REASONING_KERNEL_EVIDENCE_MATRIX.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_EVIDENCE_MATRIX.md)
23
- - [docs/ORP_REASONING_KERNEL_TECHNICAL_VALIDATION.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_TECHNICAL_VALIDATION.md)
16
+ - [docs/ORP_REASONING_KERNEL_COMPARISON_PILOT.md](./ORP_REASONING_KERNEL_COMPARISON_PILOT.md)
17
+ - [docs/ORP_REASONING_KERNEL_PICKUP_PILOT.md](./ORP_REASONING_KERNEL_PICKUP_PILOT.md)
18
+ - [docs/ORP_REASONING_KERNEL_AGENT_PILOT.md](./ORP_REASONING_KERNEL_AGENT_PILOT.md)
19
+ - [docs/ORP_REASONING_KERNEL_AGENT_REPLICATION.md](./ORP_REASONING_KERNEL_AGENT_REPLICATION.md)
20
+ - [docs/ORP_REASONING_KERNEL_CONTINUATION_PILOT.md](./ORP_REASONING_KERNEL_CONTINUATION_PILOT.md)
21
+ - [docs/ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md](./ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md)
22
+ - [docs/ORP_REASONING_KERNEL_EVIDENCE_MATRIX.md](./ORP_REASONING_KERNEL_EVIDENCE_MATRIX.md)
23
+ - [docs/ORP_REASONING_KERNEL_TECHNICAL_VALIDATION.md](./ORP_REASONING_KERNEL_TECHNICAL_VALIDATION.md)
24
24
 
25
25
  ## Evaluation Principles
26
26
 
@@ -13,13 +13,13 @@ stronger when we can say, precisely:
13
13
 
14
14
  Supporting references:
15
15
 
16
- - [docs/ORP_REASONING_KERNEL_V0_1.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_V0_1.md)
17
- - [docs/ORP_REASONING_KERNEL_TECHNICAL_VALIDATION.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_TECHNICAL_VALIDATION.md)
18
- - [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json)
19
- - [docs/ORP_REASONING_KERNEL_AGENT_PILOT.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_AGENT_PILOT.md)
20
- - [docs/ORP_REASONING_KERNEL_AGENT_REPLICATION.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_AGENT_REPLICATION.md)
21
- - [docs/ORP_REASONING_KERNEL_CONTINUATION_PILOT.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_CONTINUATION_PILOT.md)
22
- - [docs/ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md)
16
+ - [docs/ORP_REASONING_KERNEL_V0_1.md](./ORP_REASONING_KERNEL_V0_1.md)
17
+ - [docs/ORP_REASONING_KERNEL_TECHNICAL_VALIDATION.md](./ORP_REASONING_KERNEL_TECHNICAL_VALIDATION.md)
18
+ - [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](./benchmarks/orp_reasoning_kernel_v0_1_validation.json)
19
+ - [docs/ORP_REASONING_KERNEL_AGENT_PILOT.md](./ORP_REASONING_KERNEL_AGENT_PILOT.md)
20
+ - [docs/ORP_REASONING_KERNEL_AGENT_REPLICATION.md](./ORP_REASONING_KERNEL_AGENT_REPLICATION.md)
21
+ - [docs/ORP_REASONING_KERNEL_CONTINUATION_PILOT.md](./ORP_REASONING_KERNEL_CONTINUATION_PILOT.md)
22
+ - [docs/ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md](./ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md)
23
23
 
24
24
  ## Evidence Grades
25
25
 
@@ -41,24 +41,24 @@ the current kernel release.
41
41
 
42
42
  | Claim | Grade | Current Evidence | Why It Matters |
43
43
  | --- | --- | --- | --- |
44
- | ORP has a real typed kernel artifact surface. | A | [spec/v1/kernel.schema.json](/Volumes/Code_2TB/code/orp/spec/v1/kernel.schema.json), [cli/orp.py](/Volumes/Code_2TB/code/orp/cli/orp.py) | The kernel is not just prose. It is an enforceable CLI surface. |
45
- | `orp init` seeds a valid starter kernel artifact and validates it in the default flow. | A | [tests/test_orp_init.py](/Volumes/Code_2TB/code/orp/tests/test_orp_init.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json) | New repos get the kernel by default instead of needing manual adoption. |
46
- | All seven v0.1 artifact classes can scaffold and validate successfully. | A | [tests/test_orp_kernel.py](/Volumes/Code_2TB/code/orp/tests/test_orp_kernel.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json) | The kernel is broad enough for multiple project artifact types. |
47
- | Hard mode blocks invalid promotable artifacts. | A | [tests/test_orp_kernel.py](/Volumes/Code_2TB/code/orp/tests/test_orp_kernel.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json) | ORP can enforce structural promotion standards rather than only advising. |
48
- | Soft mode records invalidity without blocking work. | A | [tests/test_orp_kernel.py](/Volumes/Code_2TB/code/orp/tests/test_orp_kernel.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json) | ORP can stay fluid at intake while still surfacing missing structure. |
49
- | Existing `structure_kernel` gates remain compatible when no explicit kernel config is present. | A | [tests/test_orp_kernel.py](/Volumes/Code_2TB/code/orp/tests/test_orp_kernel.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json) | The kernel does not silently break earlier ORP configurations. |
50
- | One-shot local kernel CLI operations are within human-scale latency on the reference machine. | A | [scripts/orp-kernel-benchmark.py](/Volumes/Code_2TB/code/orp/scripts/orp-kernel-benchmark.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json) | The kernel is operationally lightweight enough to use during normal work. |
51
- | A small cross-domain reference corpus fits the current class set cleanly. | A | [examples/kernel/corpus](/Volumes/Code_2TB/code/orp/examples/kernel/corpus), [tests/test_orp_kernel_corpus.py](/Volumes/Code_2TB/code/orp/tests/test_orp_kernel_corpus.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json) | The kernel now has explicit cross-domain fit evidence, not only rationale. |
52
- | Each artifact class rejects a candidate when a required field is removed. | A | [tests/test_orp_kernel_corpus.py](/Volumes/Code_2TB/code/orp/tests/test_orp_kernel_corpus.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json) | Class-specific enforcement is directly proven instead of inferred from a subset of cases. |
53
- | The CLI validator stays aligned with the published kernel schema. | A | [tests/test_orp_kernel_corpus.py](/Volumes/Code_2TB/code/orp/tests/test_orp_kernel_corpus.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json) | The kernel no longer relies on an undocumented validator rule set drifting away from the schema. |
54
- | Equivalent YAML and JSON artifacts validate to the same semantic result. | A | [tests/test_orp_kernel_corpus.py](/Volumes/Code_2TB/code/orp/tests/test_orp_kernel_corpus.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json) | The protocol is representation-stable rather than format-sensitive. |
55
- | The validator rejects adversarial near-miss artifacts. | A | [tests/test_orp_kernel_corpus.py](/Volumes/Code_2TB/code/orp/tests/test_orp_kernel_corpus.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json) | The kernel is stronger against malformed or gameable inputs than before. |
56
- | On a matched internal comparison corpus, kernel artifacts outperform both free-form and generic checklist artifacts on structural scoring. | A | [docs/ORP_REASONING_KERNEL_COMPARISON_PILOT.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_COMPARISON_PILOT.md), [docs/benchmarks/orp_reasoning_kernel_comparison_v0_1.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_comparison_v0_1.json), [scripts/orp-kernel-comparison.py](/Volumes/Code_2TB/code/orp/scripts/orp-kernel-comparison.py) | ORP now has direct comparative evidence for structural artifact quality on a matched internal corpus, not only rationale. |
57
- | On a matched internal pickup proxy, kernel artifacts preserve more explicit handoff-critical information than both free-form and generic checklist artifacts. | A | [docs/ORP_REASONING_KERNEL_PICKUP_PILOT.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_PICKUP_PILOT.md), [docs/benchmarks/orp_reasoning_kernel_pickup_v0_1.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_pickup_v0_1.json), [scripts/orp-kernel-pickup.py](/Volumes/Code_2TB/code/orp/scripts/orp-kernel-pickup.py) | ORP now has a second comparative signal showing that kernel structure turns into more explicit pickup value, not just fuller-looking artifacts. |
58
- | On a matched live Codex recoverability simulation, kernel artifacts preserve full required-field recoverability, outperform free-form artifacts on all matched cases, and outperform generic checklist artifacts on average without per-case losses. | A | [docs/ORP_REASONING_KERNEL_AGENT_PILOT.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_AGENT_PILOT.md), [docs/benchmarks/orp_reasoning_kernel_agent_pilot_v0_1.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_agent_pilot_v0_1.json), [scripts/orp-kernel-agent-pilot.py](/Volumes/Code_2TB/code/orp/scripts/orp-kernel-agent-pilot.py) | ORP now has direct in-environment agent evidence that the kernel’s structural advantage survives contact with a real fresh downstream Codex session. |
59
- | On a `10`-repeat full-corpus live Codex replication pilot, the kernel’s recoverability advantage stays stable across fresh-session reruns, with zero invention, no run-level losses, and perfect per-field stability on required kernel fields. | A | [docs/ORP_REASONING_KERNEL_AGENT_REPLICATION.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_AGENT_REPLICATION.md), [docs/benchmarks/orp_reasoning_kernel_agent_replication_v0_2.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_agent_replication_v0_2.json), [scripts/orp-kernel-agent-replication.py](/Volumes/Code_2TB/code/orp/scripts/orp-kernel-agent-replication.py) | ORP now has stronger repeatability evidence that the live agent result is not just a single-run artifact and that the structural advantage survives at field level, not only in aggregate means. |
60
- | On a matched full-corpus live continuation pilot, kernel artifacts support the strongest continuation score, never underperform the generic checklist baseline, and keep invention at zero. | A | [docs/ORP_REASONING_KERNEL_CONTINUATION_PILOT.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_CONTINUATION_PILOT.md), [docs/benchmarks/orp_reasoning_kernel_continuation_v0_1.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_continuation_v0_1.json), [scripts/orp-kernel-continuation-pilot.py](/Volumes/Code_2TB/code/orp/scripts/orp-kernel-continuation-pilot.py) | ORP now has direct agent-first evidence that kernel artifacts are not only recoverable, but also a safe and effective base for downstream continuation. |
61
- | On a harder matched full-corpus canonical continuation pilot, kernel artifacts beat free-form on every case, beat checklist on average, and keep the lowest invention rate while revealing checklist as a real competitive baseline on some cases. | A | [docs/ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md), [docs/benchmarks/orp_reasoning_kernel_canonical_continuation_v0_1.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_canonical_continuation_v0_1.json), [scripts/orp-kernel-canonical-continuation.py](/Volumes/Code_2TB/code/orp/scripts/orp-kernel-canonical-continuation.py) | ORP now has a stricter downstream-agent benchmark where the task is not merely “continue safely,” but “produce the next canonical artifact” without inventing unsupported structure. |
44
+ | ORP has a real typed kernel artifact surface. | A | [spec/v1/kernel.schema.json](../spec/v1/kernel.schema.json), [cli/orp.py](../cli/orp.py) | The kernel is not just prose. It is an enforceable CLI surface. |
45
+ | `orp init` seeds a valid starter kernel artifact and validates it in the default flow. | A | [tests/test_orp_init.py](../tests/test_orp_init.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](./benchmarks/orp_reasoning_kernel_v0_1_validation.json) | New repos get the kernel by default instead of needing manual adoption. |
46
+ | All seven v0.1 artifact classes can scaffold and validate successfully. | A | [tests/test_orp_kernel.py](../tests/test_orp_kernel.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](./benchmarks/orp_reasoning_kernel_v0_1_validation.json) | The kernel is broad enough for multiple project artifact types. |
47
+ | Hard mode blocks invalid promotable artifacts. | A | [tests/test_orp_kernel.py](../tests/test_orp_kernel.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](./benchmarks/orp_reasoning_kernel_v0_1_validation.json) | ORP can enforce structural promotion standards rather than only advising. |
48
+ | Soft mode records invalidity without blocking work. | A | [tests/test_orp_kernel.py](../tests/test_orp_kernel.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](./benchmarks/orp_reasoning_kernel_v0_1_validation.json) | ORP can stay fluid at intake while still surfacing missing structure. |
49
+ | Existing `structure_kernel` gates remain compatible when no explicit kernel config is present. | A | [tests/test_orp_kernel.py](../tests/test_orp_kernel.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](./benchmarks/orp_reasoning_kernel_v0_1_validation.json) | The kernel does not silently break earlier ORP configurations. |
50
+ | One-shot local kernel CLI operations are within human-scale latency on the reference machine. | A | [scripts/orp-kernel-benchmark.py](../scripts/orp-kernel-benchmark.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](./benchmarks/orp_reasoning_kernel_v0_1_validation.json) | The kernel is operationally lightweight enough to use during normal work. |
51
+ | A small cross-domain reference corpus fits the current class set cleanly. | A | [examples/kernel/corpus](../examples/kernel/corpus), [tests/test_orp_kernel_corpus.py](../tests/test_orp_kernel_corpus.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](./benchmarks/orp_reasoning_kernel_v0_1_validation.json) | The kernel now has explicit cross-domain fit evidence, not only rationale. |
52
+ | Each artifact class rejects a candidate when a required field is removed. | A | [tests/test_orp_kernel_corpus.py](../tests/test_orp_kernel_corpus.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](./benchmarks/orp_reasoning_kernel_v0_1_validation.json) | Class-specific enforcement is directly proven instead of inferred from a subset of cases. |
53
+ | The CLI validator stays aligned with the published kernel schema. | A | [tests/test_orp_kernel_corpus.py](../tests/test_orp_kernel_corpus.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](./benchmarks/orp_reasoning_kernel_v0_1_validation.json) | The kernel no longer relies on an undocumented validator rule set drifting away from the schema. |
54
+ | Equivalent YAML and JSON artifacts validate to the same semantic result. | A | [tests/test_orp_kernel_corpus.py](../tests/test_orp_kernel_corpus.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](./benchmarks/orp_reasoning_kernel_v0_1_validation.json) | The protocol is representation-stable rather than format-sensitive. |
55
+ | The validator rejects adversarial near-miss artifacts. | A | [tests/test_orp_kernel_corpus.py](../tests/test_orp_kernel_corpus.py), [docs/benchmarks/orp_reasoning_kernel_v0_1_validation.json](./benchmarks/orp_reasoning_kernel_v0_1_validation.json) | The kernel is stronger against malformed or gameable inputs than before. |
56
+ | On a matched internal comparison corpus, kernel artifacts outperform both free-form and generic checklist artifacts on structural scoring. | A | [docs/ORP_REASONING_KERNEL_COMPARISON_PILOT.md](./ORP_REASONING_KERNEL_COMPARISON_PILOT.md), [docs/benchmarks/orp_reasoning_kernel_comparison_v0_1.json](./benchmarks/orp_reasoning_kernel_comparison_v0_1.json), [scripts/orp-kernel-comparison.py](../scripts/orp-kernel-comparison.py) | ORP now has direct comparative evidence for structural artifact quality on a matched internal corpus, not only rationale. |
57
+ | On a matched internal pickup proxy, kernel artifacts preserve more explicit handoff-critical information than both free-form and generic checklist artifacts. | A | [docs/ORP_REASONING_KERNEL_PICKUP_PILOT.md](./ORP_REASONING_KERNEL_PICKUP_PILOT.md), [docs/benchmarks/orp_reasoning_kernel_pickup_v0_1.json](./benchmarks/orp_reasoning_kernel_pickup_v0_1.json), [scripts/orp-kernel-pickup.py](../scripts/orp-kernel-pickup.py) | ORP now has a second comparative signal showing that kernel structure turns into more explicit pickup value, not just fuller-looking artifacts. |
58
+ | On a matched live Codex recoverability simulation, kernel artifacts preserve full required-field recoverability, outperform free-form artifacts on all matched cases, and outperform generic checklist artifacts on average without per-case losses. | A | [docs/ORP_REASONING_KERNEL_AGENT_PILOT.md](./ORP_REASONING_KERNEL_AGENT_PILOT.md), [docs/benchmarks/orp_reasoning_kernel_agent_pilot_v0_1.json](./benchmarks/orp_reasoning_kernel_agent_pilot_v0_1.json), [scripts/orp-kernel-agent-pilot.py](../scripts/orp-kernel-agent-pilot.py) | ORP now has direct in-environment agent evidence that the kernel’s structural advantage survives contact with a real fresh downstream Codex session. |
59
+ | On a `10`-repeat full-corpus live Codex replication pilot, the kernel’s recoverability advantage stays stable across fresh-session reruns, with zero invention, no run-level losses, and perfect per-field stability on required kernel fields. | A | [docs/ORP_REASONING_KERNEL_AGENT_REPLICATION.md](./ORP_REASONING_KERNEL_AGENT_REPLICATION.md), [docs/benchmarks/orp_reasoning_kernel_agent_replication_v0_2.json](./benchmarks/orp_reasoning_kernel_agent_replication_v0_2.json), [scripts/orp-kernel-agent-replication.py](../scripts/orp-kernel-agent-replication.py) | ORP now has stronger repeatability evidence that the live agent result is not just a single-run artifact and that the structural advantage survives at field level, not only in aggregate means. |
60
+ | On a matched full-corpus live continuation pilot, kernel artifacts support the strongest continuation score, never underperform the generic checklist baseline, and keep invention at zero. | A | [docs/ORP_REASONING_KERNEL_CONTINUATION_PILOT.md](./ORP_REASONING_KERNEL_CONTINUATION_PILOT.md), [docs/benchmarks/orp_reasoning_kernel_continuation_v0_1.json](./benchmarks/orp_reasoning_kernel_continuation_v0_1.json), [scripts/orp-kernel-continuation-pilot.py](../scripts/orp-kernel-continuation-pilot.py) | ORP now has direct agent-first evidence that kernel artifacts are not only recoverable, but also a safe and effective base for downstream continuation. |
61
+ | On a harder matched full-corpus canonical continuation pilot, kernel artifacts beat free-form on every case, beat checklist on average, and keep the lowest invention rate while revealing checklist as a real competitive baseline on some cases. | A | [docs/ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md](./ORP_REASONING_KERNEL_CANONICAL_CONTINUATION_PILOT.md), [docs/benchmarks/orp_reasoning_kernel_canonical_continuation_v0_1.json](./benchmarks/orp_reasoning_kernel_canonical_continuation_v0_1.json), [scripts/orp-kernel-canonical-continuation.py](../scripts/orp-kernel-canonical-continuation.py) | ORP now has a stricter downstream-agent benchmark where the task is not merely “continue safely,” but “produce the next canonical artifact” without inventing unsupported structure. |
62
62
 
63
63
  ## What Is Strong But Not Fully Sealed
64
64
 
@@ -28,8 +28,8 @@ The current core kernel remains the canonical source of truth for:
28
28
 
29
29
  Those semantics live in:
30
30
 
31
- - [spec/v1/kernel.schema.json](/Volumes/Code_2TB/code/orp/spec/v1/kernel.schema.json)
32
- - [cli/orp.py](/Volumes/Code_2TB/code/orp/cli/orp.py)
31
+ - [spec/v1/kernel.schema.json](../spec/v1/kernel.schema.json)
32
+ - [cli/orp.py](../cli/orp.py)
33
33
 
34
34
  The kernel should not self-mutate from a single chat, a single agent guess, or
35
35
  one repo’s habits.
@@ -74,7 +74,7 @@ Use it for changes like:
74
74
 
75
75
  Proposal shape is governed by:
76
76
 
77
- - [spec/v1/kernel-proposal.schema.json](/Volumes/Code_2TB/code/orp/spec/v1/kernel-proposal.schema.json)
77
+ - [spec/v1/kernel-proposal.schema.json](../spec/v1/kernel-proposal.schema.json)
78
78
 
79
79
  ### `orp kernel migrate`
80
80
 
@@ -93,7 +93,7 @@ It should begin as an extension or proposal before becoming universal.
93
93
 
94
94
  Extension shape is defined in:
95
95
 
96
- - [spec/v1/kernel-extension.schema.json](/Volumes/Code_2TB/code/orp/spec/v1/kernel-extension.schema.json)
96
+ - [spec/v1/kernel-extension.schema.json](../spec/v1/kernel-extension.schema.json)
97
97
 
98
98
  That gives ORP a place to trial domain-specific structure without forcing it
99
99
  into every project prematurely.
@@ -5,12 +5,12 @@ Reasoning Kernel.
5
5
 
6
6
  Supporting artifact:
7
7
 
8
- - [docs/benchmarks/orp_reasoning_kernel_pickup_v0_1.json](/Volumes/Code_2TB/code/orp/docs/benchmarks/orp_reasoning_kernel_pickup_v0_1.json)
8
+ - [docs/benchmarks/orp_reasoning_kernel_pickup_v0_1.json](./benchmarks/orp_reasoning_kernel_pickup_v0_1.json)
9
9
 
10
10
  Supporting corpus and harness:
11
11
 
12
- - [examples/kernel/comparison/comparison-corpus.json](/Volumes/Code_2TB/code/orp/examples/kernel/comparison/comparison-corpus.json)
13
- - [scripts/orp-kernel-pickup.py](/Volumes/Code_2TB/code/orp/scripts/orp-kernel-pickup.py)
12
+ - [examples/kernel/comparison/comparison-corpus.json](../examples/kernel/comparison/comparison-corpus.json)
13
+ - [scripts/orp-kernel-pickup.py](../scripts/orp-kernel-pickup.py)
14
14
 
15
15
  ## What This Pilot Measures
16
16
 
@@ -96,7 +96,7 @@ methodology.
96
96
  It is stronger evidence than a rationale-only claim, but it remains an
97
97
  internal, deterministic proxy. The next step after this is still a live
98
98
  human/agent pickup study as described in
99
- [docs/ORP_REASONING_KERNEL_EVALUATION_PLAN.md](/Volumes/Code_2TB/code/orp/docs/ORP_REASONING_KERNEL_EVALUATION_PLAN.md).
99
+ [docs/ORP_REASONING_KERNEL_EVALUATION_PLAN.md](./ORP_REASONING_KERNEL_EVALUATION_PLAN.md).
100
100
 
101
101
  ## Bottom Line
102
102