@riddledc/riddle-proof-packs 0.4.2 → 0.4.3

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.
Files changed (19) hide show
  1. package/README.md +2 -1
  2. package/dist/index.cjs +29 -0
  3. package/dist/index.js +29 -0
  4. package/package.json +1 -1
  5. package/packs/neon-step-sequencer/README.md +2 -1
  6. package/packs/neon-step-sequencer/case-study/findings.md +11 -0
  7. package/packs/neon-step-sequencer/case-study/ratchet-card.md +2 -1
  8. package/packs/neon-step-sequencer/case-study/ratchet-log.md +55 -0
  9. package/packs/neon-step-sequencer/case-study/reusable-lessons.md +4 -0
  10. package/packs/neon-step-sequencer/examples/README.md +2 -0
  11. package/packs/neon-step-sequencer/examples/run-006-ratchet-loop-human-review-packet/artifact-manifest.json +32 -0
  12. package/packs/neon-step-sequencer/examples/run-006-ratchet-loop-human-review-packet/console.json +4 -0
  13. package/packs/neon-step-sequencer/examples/run-006-ratchet-loop-human-review-packet/dom-summary.json +40 -0
  14. package/packs/neon-step-sequencer/examples/run-006-ratchet-loop-human-review-packet/profile-result.json +11786 -0
  15. package/packs/neon-step-sequencer/examples/run-006-ratchet-loop-human-review-packet/proof.json +11786 -0
  16. package/packs/neon-step-sequencer/examples/run-006-ratchet-loop-human-review-packet/screenshots/lilarcade-neon-ratchet-loop-mix-level-search-desktop-neon-ratchet-loop-mix-level-search.png +0 -0
  17. package/packs/neon-step-sequencer/examples/run-006-ratchet-loop-human-review-packet/screenshots/lilarcade-neon-ratchet-loop-mix-level-search-desktop.png +0 -0
  18. package/packs/neon-step-sequencer/examples/run-006-ratchet-loop-human-review-packet/summary.md +44 -0
  19. package/packs/neon-step-sequencer/profiles/ratchet-loop-mix-level-search.json +29 -0
package/README.md CHANGED
@@ -90,12 +90,13 @@ Profiles are stored under `packs/<slug>/profile.json` and mirrored into the runt
90
90
  - `neon-step-sequencer-mobile-trainer-layout`
91
91
  - `neon-step-sequencer-full-mix-health-matrix`
92
92
  - `neon-step-sequencer-explore-songs-and-mixes`
93
+ - `neon-step-sequencer-ratchet-loop-mix-level-search`
93
94
 
94
95
  ## Audio and Neon ratchet packs
95
96
 
96
97
  The `audio-mix` directory contains reusable audio-proof authoring guidance, a profile template, a metrics schema, a ratchet method, and a human-review rubric.
97
98
 
98
- The `neon-step-sequencer` directory contains the first app-specific ratchet lab under the new architecture. Its profiles declare `current_target` or `interaction_snapshots` evidence-role patterns and explicitly state what they do not prove. The case-study files record the claim, evidence, failure classification, smallest layer changed, and next sharper question for each run.
99
+ The `neon-step-sequencer` directory contains the first app-specific ratchet lab under the new architecture. Its profiles declare `current_target` or `interaction_snapshots` evidence-role patterns and explicitly state what they do not prove. The ratchet-loop profile now expects a compact `humanReviewPacket` for listening handoff: supported/rejected candidates, objective guardrails, state restoration, review-order ranking, and taste caveats. The case-study files record the claim, evidence, failure classification, smallest layer changed, and next sharper question for each run.
99
100
 
100
101
  ## Usage
101
102
 
package/dist/index.cjs CHANGED
@@ -1695,6 +1695,7 @@ var ratchet_loop_mix_level_search_default = {
1695
1695
  path: "__NEON_MIX_PROOF__.runRatchetLoop",
1696
1696
  args: [
1697
1697
  {
1698
+ intent: "turn the chord part down a little",
1698
1699
  strategy: "mix-level-search",
1699
1700
  focusTracks: [
1700
1701
  "bass",
@@ -1729,6 +1730,15 @@ var ratchet_loop_mix_level_search_default = {
1729
1730
  },
1730
1731
  {
1731
1732
  path: "supportedClaimCandidateCount"
1733
+ },
1734
+ {
1735
+ path: "humanReviewPacket.status"
1736
+ },
1737
+ {
1738
+ path: "humanReviewPacket.recommendation.candidate.action.track"
1739
+ },
1740
+ {
1741
+ path: "humanReviewPacket.ranking.role"
1732
1742
  }
1733
1743
  ]
1734
1744
  },
@@ -1738,6 +1748,24 @@ var ratchet_loop_mix_level_search_default = {
1738
1748
  expected_value: true,
1739
1749
  timeout_ms: 1e4
1740
1750
  },
1751
+ {
1752
+ type: "assert_window_value",
1753
+ path: "__neonMixProof.ratchetLoop.humanReviewPacket.kind",
1754
+ expected_value: "human_review_packet",
1755
+ timeout_ms: 1e4
1756
+ },
1757
+ {
1758
+ type: "assert_window_value",
1759
+ path: "__neonMixProof.ratchetLoop.humanReviewPacket.ranking.role",
1760
+ expected_value: "review_order_only",
1761
+ timeout_ms: 1e4
1762
+ },
1763
+ {
1764
+ type: "assert_window_value",
1765
+ path: "__neonMixProof.ratchetLoop.humanReviewPacket.request.candidateActionsAreTransient",
1766
+ expected_value: true,
1767
+ timeout_ms: 1e4
1768
+ },
1741
1769
  {
1742
1770
  type: "screenshot",
1743
1771
  label: "neon-ratchet-loop-mix-level-search",
@@ -1785,6 +1813,7 @@ var ratchet_loop_mix_level_search_default = {
1785
1813
  "claim candidates are captured",
1786
1814
  "each candidate records claim receipts",
1787
1815
  "best supported claim candidate or human-review status is captured",
1816
+ "compact human-review packet is captured",
1788
1817
  "state restoration receipt is captured"
1789
1818
  ],
1790
1819
  does_not_prove: [
package/dist/index.js CHANGED
@@ -1660,6 +1660,7 @@ var ratchet_loop_mix_level_search_default = {
1660
1660
  path: "__NEON_MIX_PROOF__.runRatchetLoop",
1661
1661
  args: [
1662
1662
  {
1663
+ intent: "turn the chord part down a little",
1663
1664
  strategy: "mix-level-search",
1664
1665
  focusTracks: [
1665
1666
  "bass",
@@ -1694,6 +1695,15 @@ var ratchet_loop_mix_level_search_default = {
1694
1695
  },
1695
1696
  {
1696
1697
  path: "supportedClaimCandidateCount"
1698
+ },
1699
+ {
1700
+ path: "humanReviewPacket.status"
1701
+ },
1702
+ {
1703
+ path: "humanReviewPacket.recommendation.candidate.action.track"
1704
+ },
1705
+ {
1706
+ path: "humanReviewPacket.ranking.role"
1697
1707
  }
1698
1708
  ]
1699
1709
  },
@@ -1703,6 +1713,24 @@ var ratchet_loop_mix_level_search_default = {
1703
1713
  expected_value: true,
1704
1714
  timeout_ms: 1e4
1705
1715
  },
1716
+ {
1717
+ type: "assert_window_value",
1718
+ path: "__neonMixProof.ratchetLoop.humanReviewPacket.kind",
1719
+ expected_value: "human_review_packet",
1720
+ timeout_ms: 1e4
1721
+ },
1722
+ {
1723
+ type: "assert_window_value",
1724
+ path: "__neonMixProof.ratchetLoop.humanReviewPacket.ranking.role",
1725
+ expected_value: "review_order_only",
1726
+ timeout_ms: 1e4
1727
+ },
1728
+ {
1729
+ type: "assert_window_value",
1730
+ path: "__neonMixProof.ratchetLoop.humanReviewPacket.request.candidateActionsAreTransient",
1731
+ expected_value: true,
1732
+ timeout_ms: 1e4
1733
+ },
1706
1734
  {
1707
1735
  type: "screenshot",
1708
1736
  label: "neon-ratchet-loop-mix-level-search",
@@ -1750,6 +1778,7 @@ var ratchet_loop_mix_level_search_default = {
1750
1778
  "claim candidates are captured",
1751
1779
  "each candidate records claim receipts",
1752
1780
  "best supported claim candidate or human-review status is captured",
1781
+ "compact human-review packet is captured",
1753
1782
  "state restoration receipt is captured"
1754
1783
  ],
1755
1784
  does_not_prove: [
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@riddledc/riddle-proof-packs",
3
- "version": "0.4.2",
3
+ "version": "0.4.3",
4
4
  "description": "Reusable proof pack profiles and metadata helpers for the Riddle proof framework.",
5
5
  "license": "MIT",
6
6
  "author": "RiddleDC",
@@ -33,7 +33,7 @@ This pack is the first app-specific lab for the open Riddle Proof architecture.
33
33
 
34
34
  The loop is not mix-specific as a proof concept. The proof concept is a bounded ratchet loop: propose a candidate, apply it, collect evidence, classify the result, restore or keep state, and repeat until the budget is exhausted.
35
35
 
36
- This pack's first concrete strategy is `mix-level-search`, which turns small level edits into change-claim candidates. Each candidate says what action will be attempted, what receipts must support the claim, and what evidence should be reviewed afterward. The loop may include a ranking metric to order review, but the verdict comes from receipts and invariants, not from a universal mix-quality number. It still does not decide subjective mix taste; the output is a listening and review packet.
36
+ This pack's first concrete strategy is `mix-level-search`, which turns small level edits into change-claim candidates. Each candidate says what action will be attempted, what receipts must support the claim, and what evidence should be reviewed afterward. The loop may include a ranking metric to order review, but the verdict comes from receipts and invariants, not from a universal mix-quality number. It still does not decide subjective mix taste; the output is a `humanReviewPacket` for listening handoff.
37
37
 
38
38
  ## Example evidence
39
39
 
@@ -44,6 +44,7 @@ The `examples/` directory contains local Playwright proof results captured again
44
44
  - `run-003-full-matrix`: passing `current_target` viewport matrix across desktop, phone, iPad Mini, and iPad with `0 px` horizontal overflow.
45
45
  - `run-004-ratchet-loop-mix-level-search`: passing `interaction_snapshots` proof where a bounded loop tested six mix-level change-claim candidates, found a supported `chord -0.10` candidate, recorded receipt-level verdicts, and restored app state without keeping the edit.
46
46
  - `run-005-explore-songs-and-mixes-final`: passing `current_target` exploration sweep across four songs and eight song/part entries, with `8` passing entries, `0` prioritized findings, and no clipping after the local app-contract and mix-headroom ratchet.
47
+ - `run-006-ratchet-loop-human-review-packet`: passing `interaction_snapshots` proof where the same bounded loop returned a compact `humanReviewPacket` with the recommended `chord -0.10` candidate, objective guardrails, `review_order_only` ranking, state restoration, and explicit listening caveats.
47
48
 
48
49
  ## Naming note
49
50
 
@@ -112,6 +112,17 @@ Pack summary guidance first; Riddle Proof core only if a general display primiti
112
112
  - rerun: passed on May 24, 2026 with `8` entries, `8` passed, and `0` findings.
113
113
  - next sharper question: can this exploration workflow become the normal local pack/profile loop before any changeset or npm release?
114
114
 
115
+ ### Run 006 made the listening handoff compact
116
+
117
+ - run: `run-006-ratchet-loop-human-review-packet`
118
+ - claim: a bounded ratchet loop can return a compact human-review packet that summarizes supported candidates, objective guardrails, state restoration, ranking-as-review-order, and listening caveats.
119
+ - observed evidence: loop status was `claim_candidate_supported`; the packet kind was `human_review_packet`; packet status was `candidate_ready_for_listening_review`; the recommended candidate was `chord -0.10`; supported candidate count was `6`; rejected candidate count was `0`; ranking role was `review_order_only`; app state was restored after the loop; permanent edit was not kept.
120
+ - classification: none; passing `interaction_snapshots` proof with subjective listening caveat.
121
+ - smallest layer changed: app proof contract and proof-pack profile assertions.
122
+ - change made: added `humanReviewPacket` to the Neon ratchet-loop result and updated the pack profile to assert the packet shape.
123
+ - rerun: passed on May 24, 2026 with local Playwright.
124
+ - next sharper question: can the packet become the standard output shape for one-off and background candidate operators across more than `mix-level-search`?
125
+
115
126
  ### Local runner shutdown needs a small ergonomics follow-up
116
127
 
117
128
  - run: `run-002-mix-change`, `run-003-full-matrix`, `run-004-ratchet-loop-mix-level-search`
@@ -35,7 +35,8 @@ The project shows that a complex audio app can improve proof confidence mostly b
35
35
  - Run 003: a `current_target` matrix passed across desktop, phone, iPad Mini, and iPad with `0 px` horizontal overflow.
36
36
  - Run 004: a bounded loop tested six `mix-level-search` change-claim candidates, returned `chord -0.10` as a supported review candidate with receipt-level evidence, and restored app state.
37
37
  - Run 005: a `current_target` exploration sweep sampled four songs and eight song/part entries. The first real sweep exposed app-contract normalization gaps and hot built-in song presets; after the local ratchet fixes it passed with `8` entries, `0` findings, and no clipping.
38
+ - Run 006: the bounded loop returned a compact `humanReviewPacket` with the supported `chord -0.10` candidate, objective guardrails, state restoration, review-order ranking, and listening caveats.
38
39
 
39
40
  ## Honest boundary
40
41
 
41
- These runs prove objective claims about a running app target. They do not prove that the mix is tasteful, that every song section is healthy, or that a release candidate is better than production. Run 005 is still bounded to configured song/part limits. The ratchet loop is a generic proof-loop shape; `mix-level-search` is only the first Neon strategy plugged into it.
42
+ These runs prove objective claims about a running app target. They do not prove that the mix is tasteful, that every song section is healthy, or that a release candidate is better than production. Run 005 is still bounded to configured song/part limits. Run 006 makes the human handoff explicit: the packet is a compact review object, not an automated taste verdict. The ratchet loop is a generic proof-loop shape; `mix-level-search` is only the first Neon strategy plugged into it.
@@ -282,6 +282,61 @@ Next sharper question:
282
282
 
283
283
  Can the same exploration shape become a reusable pack/profile workflow where a user can choose a bounded target set, run locally during iteration, and publish only after the evidence is worth sharing?
284
284
 
285
+ ## Run 006 - Human-review packet handoff
286
+
287
+ Claim:
288
+
289
+ Neon can turn the bounded `mix-level-search` loop into a compact handoff object that a human or follow-on agent can review without reading the full proof JSON.
290
+
291
+ Profile:
292
+
293
+ `profiles/ratchet-loop-mix-level-search.json`
294
+
295
+ Evidence to capture:
296
+
297
+ - loop status and supported candidate count
298
+ - recommended candidate action
299
+ - per-candidate objective guardrail summary
300
+ - state restoration receipt
301
+ - ranking role as review order only
302
+ - listening-review caveats
303
+
304
+ Possible outcomes:
305
+
306
+ - `claim_candidate_supported`: at least one candidate has the receipts needed to support its change claim and the packet can recommend it for listening review.
307
+ - `needs_human_review`: evidence is valid but no candidate satisfies every objective receipt.
308
+ - `proof_insufficient`: the app contract does not expose a packet or enough candidate receipts.
309
+ - `profile_calibration`: the packet exists but points to the wrong window, target, or candidate set.
310
+
311
+ Observed status:
312
+
313
+ Passed on May 24, 2026 with `local-playwright`.
314
+
315
+ Observed evidence:
316
+
317
+ - loop status `claim_candidate_supported`
318
+ - packet kind `human_review_packet`
319
+ - packet status `candidate_ready_for_listening_review`
320
+ - recommended candidate `chord -0.10`
321
+ - recommended action `set_mixer_level chord 0.38 -> 0.28`
322
+ - supported candidates `6`
323
+ - rejected candidates `0`
324
+ - ranking role `review_order_only`
325
+ - state restored after loop `true`
326
+ - permanent edit kept `false`
327
+
328
+ Failure classification:
329
+
330
+ None. This was a passing `interaction_snapshots` proof with an explicit listening-review caveat.
331
+
332
+ Smallest layer changed:
333
+
334
+ App proof contract and proof-pack profile. Riddle Proof core did not need a change.
335
+
336
+ Next sharper question:
337
+
338
+ Can one-off commands and background runs use this packet as their common output surface while strategy-specific code remains behind the app contract?
339
+
285
340
  ## Project note
286
341
 
287
342
  The ratchet is not a pass. The ratchet is the next sharper question.
@@ -22,6 +22,10 @@ Large metrics belong in artifacts. The summary should answer:
22
22
  - did the intended metric move?
23
23
  - what should a human review next?
24
24
 
25
+ ## Handoff packets beat raw JSON spelunking
26
+
27
+ The full proof result should remain auditable, but one-off and background loops need a compact handoff object. A good `humanReviewPacket` lists supported and rejected candidates, objective guardrails, restoration status, review-order ranking, and caveats that separate proof from taste.
28
+
25
29
  ## Core changes are last
26
30
 
27
31
  Most ratchet steps should change profile JSON, pack docs, app proof contracts, or app fixtures. Riddle Proof core changes are justified only when the missing primitive applies beyond Neon.
@@ -13,6 +13,7 @@ The raw `profile-result.json` files are real runner outputs. They intentionally
13
13
  | `run-003-full-matrix` | `current_target` | passed | The mix-health proof holds across desktop, phone, iPad Mini, and iPad viewports. |
14
14
  | `run-004-ratchet-loop-mix-level-search` | `interaction_snapshots` | passed | A bounded ratchet loop tests mix-level change-claim candidates, records receipt verdicts, chooses a supported candidate for review, and restores app state. |
15
15
  | `run-005-explore-songs-and-mixes-final` | `current_target` | passed | A bounded exploration sweep samples four songs and eight song/part entries, producing a zero-finding confidence map after app-contract and mix-headroom fixes. |
16
+ | `run-006-ratchet-loop-human-review-packet` | `interaction_snapshots` | passed | A bounded ratchet loop returns a compact `humanReviewPacket` for handoff: supported candidates, objective guardrails, state restoration, review-order ranking, and listening caveats. |
16
17
 
17
18
  ## What these examples do not prove
18
19
 
@@ -22,3 +23,4 @@ The raw `profile-result.json` files are real runner outputs. They intentionally
22
23
  - They do not prove a reference/candidate release delta. The mix-change run uses pre-action/post-action snapshots inside one proof run, not a separate baseline deployment.
23
24
  - The ratchet-loop run does not prove that the loop primitive is mix-specific; `mix-level-search` is only this pack's first concrete strategy.
24
25
  - The ratchet-loop run does not prove that the supported candidate should be kept. Its ranking metric is a review-order hint, not a taste verdict.
26
+ - The human-review packet does not replace listening judgment. It compresses objective receipts and caveats so a person or follow-on agent can decide what to review next.
@@ -0,0 +1,32 @@
1
+ {
2
+ "version": "riddle-proof-local-runner-manifest.v1",
3
+ "runner": "local-playwright",
4
+ "profile_name": "lilarcade-neon-ratchet-loop-mix-level-search",
5
+ "captured_at": "2026-05-24T06:15:38.159Z",
6
+ "artifacts": [
7
+ {
8
+ "name": "proof.json",
9
+ "path": "proof.json",
10
+ "kind": "text",
11
+ "bytes": 497403
12
+ },
13
+ {
14
+ "name": "console.json",
15
+ "path": "console.json",
16
+ "kind": "text",
17
+ "bytes": 38
18
+ },
19
+ {
20
+ "name": "dom-summary.json",
21
+ "path": "dom-summary.json",
22
+ "kind": "text",
23
+ "bytes": 815
24
+ },
25
+ {
26
+ "name": "profile-result.json",
27
+ "path": "profile-result.json",
28
+ "kind": "text",
29
+ "bytes": 497403
30
+ }
31
+ ]
32
+ }
@@ -0,0 +1,40 @@
1
+ {
2
+ "expected_viewport_count": 1,
3
+ "viewport_count": 1,
4
+ "partial": false,
5
+ "routes": [
6
+ {
7
+ "requested": "http://127.0.0.1:5177/games/drum-sequencer?song=monkberry-moon-delight-tab&mix=profile&view=trainer&instrument=bass",
8
+ "observed": "/games/drum-sequencer",
9
+ "expected_path": "/games/drum-sequencer",
10
+ "matched": true,
11
+ "http_status": 200
12
+ }
13
+ ],
14
+ "titles": [
15
+ "LilArcade"
16
+ ],
17
+ "overflow_px": [
18
+ 0
19
+ ],
20
+ "bounds_overflow_px": [
21
+ 0
22
+ ],
23
+ "overflow_offender_counts": [
24
+ 0
25
+ ],
26
+ "frames": [
27
+ {
28
+ "viewport": "desktop",
29
+ "selectors": []
30
+ }
31
+ ],
32
+ "http_status": [],
33
+ "link_status": [],
34
+ "route_inventory": [],
35
+ "network_mock_count": 0,
36
+ "network_mock_hit_count": 0,
37
+ "dialog_count": 0,
38
+ "dialog_accept_count": 0,
39
+ "dialog_dismiss_count": 0
40
+ }