Amend CB-EV-0001; add CB-WP-0002 for cost accounting

AM-4: measured what each remediation option actually buys, rather than
leaving one recommendation unquantified. serde_yaml optional removes 6
crates, not 5 — ryu belongs to that group, since serde_json now uses
zmij for floats. Full ladder: -6 to 27, serde_json -4 more to 23,
inlining SHA-256 -8 to 19, inlining ChaCha12 -4 to 15. Only
reimplementing a primitive gets under 20, so the target is unreachable
without undoing K5/K7.

Also records that crate count compares badly across ecosystems, and
offers the alternative the count is a proxy for: 307,317 lines of
third-party source under audit against 3,398 of our own.

AM-12: corrected from "uncomputable" to measured. The refusal to
estimate was right; the claim that no instrument existed was wrong.
Session transcripts carry exact per-message usage including the cache
breakdown. This session cost $248.46 at Fable 5 rates, of which 53% is
cache reads — cost is driven by context size times turn count, not by
output volume. What is still missing is per-task attribution, since
nothing marks task boundaries in a transcript.

CB-WP-0002 makes cost measurable and attributable: survey the
instruments, decide the attribution model by ADR, spec metrics that
include cost composition rather than a bare total, build a collector
whose positive control refuses to emit unreconciled numbers, and prove
it by answering a question that could not be answered before.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

.claude/ralph-loop.local.md

@@ -1,21 +0,0 @@
----
-active: true
-iteration: 3
-session_id: 8cbd5701-a096-45a4-a419-9b7b1c9419bc
-max_iterations: 20
-completion_promise: "HEUREKA"
-workplan_id: CB-WP-0001
-workplan_file: workplans/CB-WP-0001-inner-loop.md
-started_at: "2026-07-31T00:08:23Z"
----
-
-Read the workplan at `workplans/CB-WP-0001-inner-loop.md`.
-
-If every task has `status: done` AND frontmatter `status: done`:
-run `rm -f .claude/ralph-loop.local.md` first (deactivates the loop so the stop hook exits cleanly),
-then output <promise>HEUREKA</promise>.
-
-Otherwise implement the next `todo` task as described in the workplan.
-Set task `in_progress` when starting, `done` when complete.
-When all tasks are done set frontmatter `status: done`.
-

WORK-RECORDS.md

@@ -8,13 +8,13 @@
 
 | Kind | ID | Status | Lane | Source |
 | --- | --- | --- | --- | --- |
-| workplan | CB-WP-0001 | active | — | workplans/CB-WP-0001-inner-loop.md |
+| workplan | CB-WP-0001 | done | — | workplans/CB-WP-0001-inner-loop.md |
 | task | CB-WP-0001-T01 | done | — | workplans/CB-WP-0001-inner-loop.md |
 | task | CB-WP-0001-T02 | done | — | workplans/CB-WP-0001-inner-loop.md |
-| task | CB-WP-0001-T03 | todo | — | workplans/CB-WP-0001-inner-loop.md |
-| task | CB-WP-0001-T04 | todo | — | workplans/CB-WP-0001-inner-loop.md |
-| task | CB-WP-0001-T05 | todo | — | workplans/CB-WP-0001-inner-loop.md |
-| task | CB-WP-0001-T06 | todo | — | workplans/CB-WP-0001-inner-loop.md |
-| task | CB-WP-0001-T07 | todo | — | workplans/CB-WP-0001-inner-loop.md |
-| task | CB-WP-0001-T08 | todo | — | workplans/CB-WP-0001-inner-loop.md |
-| task | CB-WP-0001-T09 | todo | — | workplans/CB-WP-0001-inner-loop.md |
+| task | CB-WP-0001-T03 | done | — | workplans/CB-WP-0001-inner-loop.md |
+| task | CB-WP-0001-T04 | done | — | workplans/CB-WP-0001-inner-loop.md |
+| task | CB-WP-0001-T05 | done | — | workplans/CB-WP-0001-inner-loop.md |
+| task | CB-WP-0001-T06 | done | — | workplans/CB-WP-0001-inner-loop.md |
+| task | CB-WP-0001-T07 | done | — | workplans/CB-WP-0001-inner-loop.md |
+| task | CB-WP-0001-T08 | done | — | workplans/CB-WP-0001-inner-loop.md |
+| task | CB-WP-0001-T09 | done | — | workplans/CB-WP-0001-inner-loop.md |

evidence/CB-EV-0001-game-kernel.md

@@ -1,7 +1,9 @@
 # CB-EV-0001 — GROUND game kernel: acceptance evidence
 
 Status: **T08 complete, with one acceptance metric not met (AM-4).**
-Recorded: 2026-07-31
+Recorded: 2026-07-31. Amended 2026-07-31 — §4 gains measured savings per
+remediation option, and §5 corrects AM-12 from "uncomputable" to
+measured-at-session-level; see CB-WP-0002.
 Workplan: CB-WP-0001, task T08
 Spec: `specs/GameKernel.md` §4 (AM-1..AM-12)
 Baseline: `research/CB-RES-0001-game-kernel.md`, measurements in
@@ -23,7 +25,7 @@ Machine: WSL2, Linux 6.18.33.2-microsoft-standard-WSL2, rustc 1.97.1,
 | AM-8 lint | fmt + clippy clean | clean, `-D warnings` | **met** |
 | AM-10 foreign types | zero `HashMap`/`HashSet` | 0 | **met** |
 | AM-11 impl pairs | null + reference per port | 1 of 1 (`KernelRng`) | **met, narrow** |
-| AM-12 cost log | present | §5 | **met** |
+| AM-12 cost | per-task USD | $248.46 session; per-task pending | **partial** |
 
 AM-2, AM-3, AM-5 and AM-9 are not reported: see §6.
 
@@ -121,8 +123,8 @@ Attribution:
 |---|---|---|
 | `sha2` (K7 state hashing) | sha2, digest, block-buffer, crypto-common, generic-array, typenum, cpufeatures, cfg-if | 8 |
 | serde derive chain | serde_derive, proc-macro2, quote, syn, unicode-ident | 5 |
-| serde runtime + json | serde, serde_core, serde_json, itoa, ryu, memchr, zmij | 7 |
-| `serde_yaml` (scenario files only) | serde_yaml, unsafe-libyaml, indexmap, hashbrown, equivalent | 5 |
+| serde runtime + json | serde, serde_core, serde_json, itoa, memchr, zmij | 6 |
+| `serde_yaml` (scenario files only) | serde_yaml, unsafe-libyaml, indexmap, hashbrown, equivalent, ryu | 6 |
 | `rand_chacha` (K5 seeded RNG) | rand_chacha, rand_core, ppv-lite86, zerocopy | 4 |
 | Clay-Borg crates | cb-kernel, cb-events, cb-game-runtime, games-ground | 4 |
 
@@ -130,20 +132,31 @@ The honest options, in order of preference:
 
 1. **Make `serde_yaml` optional** behind a `scenarios` feature. YAML is
    a test-and-tooling concern; a shipped game runtime does not need it.
-   Removes 5 crates from the default build for no loss of capability.
+   Removes 6 crates from the default build for no loss of capability
+   (`ryu` belongs to this group, not to serde_json, which uses `zmij`
+   for floats — corrected after measuring the reverse-dependency graph).
    This is the one to do first, and it improves D4 optionality as well
    as D2.
-2. **Revisit the target.** ≤20 was set before the K5/K7 contracts named
-   ChaCha and SHA-256. Those two contracts cost 12 crates between them
-   and are load-bearing for determinism. A target that a spec's own
-   contracts make unreachable is a bad target.
+2. **Revisit the target, and what it measures.** Measured savings per
+   option: serde_yaml optional −6 (→27); replacing serde_json −4 more
+   (→23); inlining SHA-256 −8 (→19); inlining ChaCha12 −4 (→15). **Only
+   reimplementing SHA-256 or ChaCha gets under 20**, so the target is
+   unreachable without undoing K5/K7.
+
+   Crate count also compares badly across ecosystems: Rust splits
+   crates far more finely than npm, so "33 vs 120 npm packages" flatters
+   us. The measurable thing crate count proxies for is third-party
+   source under audit: **307,317 lines** across all five groups, against
+   3,398 of our own. Retargeting AM-4 on audited third-party LOC, split
+   into shipped-runtime and dev-toolchain, measures the real concern and
+   cannot be gamed by crate granularity.
 
 What we are **not** doing: hand-rolling SHA-256 or ChaCha to win a
 dependency count. That trades an auditable, well-tested primitive for a
 number on a scoreboard.
 
-This is a T09 input: either the metric moves for a stated reason, or
-option 1 lands and the remainder is justified.
+Carried forward as an open decision (see the note at the head of this
+file): AM-4 is re-measured once the option is chosen.
 
 ## 5. Cost log (AM-12)
 
@@ -152,14 +165,30 @@ Per `specs/MetricsAndScenarios.md` §1a. Model: Claude Fable 5, at
 
 | Task | Model | Iterations | Notes |
 |---|---|---|---|
-| T08 | claude-fable-5 | 6 code iterations + benchmarks | Token counts not captured per iteration; see limitation below |
+| CB-WP-0001 (whole session, T01–T09) | claude-fable-5 | 6 T08 code iterations + benchmarks | $248.46 measured; per-task split pending CB-WP-0002 |
 
-**Limitation, stated rather than fabricated:** exact per-task token
-counts were not instrumented during T08, so the USD figure the metric
-asks for cannot be computed honestly from this run. Recording an
-estimate here would defeat the purpose of the metric. T09 should either
-wire real token accounting into the loop or drop M-D2-CST as
-unmeasurable in this setup.
+**Correction (2026-07-31).** This section originally recorded AM-12 as
+*uncomputable*. That was wrong. The declining to estimate was right; the
+conclusion that no instrument existed was not. Every session transcript
+(`~/.claude/projects/<slug>/<session>.jsonl`) carries exact per-message
+`usage` including the cache breakdown. Read for this session:
+
+| Component | Tokens | Cost (Fable 5) |
+|---|---|---|
+| Output | 585,528 | $29.28 |
+| Cache read | 131,863,164 | $131.86 |
+| Cache write (1h) | 4,365,668 | $87.31 |
+| Input | 1,090 | $0.01 |
+| **Session total** | | **$248.46** (~$124 on Opus 5) |
+
+**53% of the cost is cache reads**, not output. Cost in an agentic loop
+is driven by context size × turn count, which no "tokens per task"
+metric would have surfaced.
+
+Still missing is *attribution*: this is a whole-session figure, not a
+per-task one, because nothing marks task boundaries in the transcript.
+That is what CB-WP-0002 is for. The AM-12 row above should be read as
+"session-level cost measured; per-task attribution pending CB-WP-0002".
 
 ## 6. Metrics not reported
 

workplans/CB-WP-0002-cost-accounting.md

@@ -0,0 +1,185 @@
+---
+id: CB-WP-0002
+title: "Make agentic cost measurable, so D2 claims are falsifiable"
+status: proposed
+---
+
+# Purpose
+
+CB-WP-0001 specified a cost metric (AM-12, `M-D2-CST`) in full — price
+sheet, formula, staleness rule — and then could not report a number,
+because nothing was ever instrumented. It was recorded as
+"uncomputable".
+
+That was wrong, and the retrospective for this workplan should say so
+plainly: **the data existed the whole time.** Every Claude Code session
+transcript (`~/.claude/projects/<slug>/<session>.jsonl`) carries exact
+per-message `usage`, including the cache breakdown. Reading it for
+CB-WP-0001's session gives:
+
+| Component | Tokens | Cost (Fable 5) |
+|---|---|---|
+| Output | 585,528 | $29.28 |
+| Cache read | 131,863,164 | **$131.86** |
+| Cache write (1h) | 4,365,668 | $87.31 |
+| Input | 1,090 | $0.01 |
+| **Total** | | **$248.46** |
+
+The headline finding from that single reading is the reason this
+workplan exists: **53% of the cost is cache reads**, not output. Cost in
+an agentic loop is driven by context size × turn count, not by how much
+the model writes. No D2 decision made on "tokens per task" would have
+surfaced that.
+
+This workplan makes cost a first-class measured dimension so that
+future AM-12 equivalents support conclusions instead of decorating an
+evidence file. It follows InnerLoop v1.0, including the rules that pass
+earned: every metric names its instrument, and every harness carries a
+positive control.
+
+## Phase A — Instrument
+
+## Task: Survey the available cost instruments
+
+```task
+id: CB-WP-0002-T01
+status: todo
+priority: high
+```
+
+Produce `research/CB-RES-0002-cost-accounting.md` per the InnerLoop
+survey template. Candidates at minimum: session transcript JSONL
+(`usage` per assistant message), the Custodian State Hub token-event
+API (`record_token_event`, `update_task_status` token tiers,
+`get_token_summary`), the Claude Code status bar, and the Anthropic
+usage/billing API. Per candidate: what it reports, granularity,
+accuracy, whether it survives compaction, and whether it can attribute
+cost to a unit of work.
+
+Note explicitly which are **exact** and which are **estimates** — the
+prior failure was tolerating an estimate-shaped metric. Name the
+benchmark-to-beat per dimension; expect the transcript to lead on
+accuracy and the hub to lead on durability.
+
+## Task: Decide the instrument and the attribution model (ADR)
+
+```task
+id: CB-WP-0002-T02
+status: todo
+priority: high
+```
+
+Adversarial review of T01 first (InnerLoop §Step 2), committed as
+`history/YYMMDD-cost-accounting-{challenge,response}.md`.
+
+Then `decisions/ADR-0003-cost-accounting.md`. The genuinely hard part is
+**attribution**: a transcript is a flat message stream, and a "task" is
+a workplan concept. Options to weigh, not assume:
+
+- git commit timestamps as task boundaries (the loop commits per
+  iteration, so boundaries already exist and are durable)
+- explicit session markers emitted at task start/end
+- hub task status transitions as the time index
+
+State the expected advantage per dimension and the known failure modes
+of the chosen model — in particular, what happens to attribution across
+`/compact`, across resumed sessions, and for work spanning a boundary.
+Gate: no collector code before this ADR is committed.
+
+## Task: Specify the cost metrics with named instruments
+
+```task
+id: CB-WP-0002-T03
+status: todo
+priority: high
+```
+
+Write `specs/CostAccounting.md`: the cost model (input, output, cache
+read, cache write at 5m and 1h, per the price sheet), the attribution
+contract, and the acceptance metrics — each naming the **command that
+produces its number**, per InnerLoop v1.0 §Step 4.
+
+Must include a metric for the finding that motivated this workplan:
+**cost composition** (what fraction is cache read vs write vs output),
+not only a total. A single total would have hidden the 53%.
+
+Revise `specs/MetricsAndScenarios.md` §1a to point at this spec, and
+replace AM-12's definition with one that is computable.
+
+## Phase B — Build and prove
+
+## Task: Implement the cost collector
+
+```task
+id: CB-WP-0002-T04
+status: todo
+priority: medium
+```
+
+Implement the tool chosen in T02 (expected: `tools/cb-cost`). It reads
+transcripts, applies the price sheet at
+`benchmarks/baselines/model-prices.toml`, attributes cost per the T02
+model, and emits both a per-task table and a composition breakdown in
+the evidence-row format from T03.
+
+**Positive control is mandatory** (InnerLoop v1.0 §Step 5): the tool
+asserts that attributed tokens sum to the transcript total, and refuses
+to emit numbers when they do not reconcile. An unattributed remainder is
+reported as its own line, never silently dropped — the failure this
+whole workplan exists to prevent was a number that looked fine.
+
+Handle multi-model sessions: a session that switches models must price
+each message at its own model's rate.
+
+## Task: Validate against CB-WP-0001 and answer a real question
+
+```task
+id: CB-WP-0002-T05
+status: todo
+priority: medium
+```
+
+Run the collector over the CB-WP-0001 session and commit
+`evidence/CB-EV-0002-cost-accounting.md`. Reconciliation against the
+totals in this workplan's Purpose section is the acceptance test: the
+tool must reproduce $248.46 (Fable 5) from the same transcript, or
+explain the difference.
+
+Then use it to answer at least one question that could not be answered
+before, and record the answer. Candidates: which of T01–T09 cost most
+and whether that matched its value; what a `/compact` costs; whether the
+adversarial review paid for itself; how much the six T08 code iterations
+cost relative to the research phase.
+
+A tool that produces numbers nobody draws a conclusion from has not
+cleared the bar that CB-WP-0001's AM-12 failed to clear.
+
+## Task: Wire cost into the loop
+
+```task
+id: CB-WP-0002-T06
+status: todo
+priority: low
+```
+
+Make cost collection automatic rather than remembered: add the
+collector to the evidence checklist in `specs/InnerLoop.md`, emit hub
+token events so the Token Cost dashboard reflects real numbers rather
+than the 1000/500 heuristic fallback, and add a `make cost` target to
+the one command surface.
+
+## Task: Retrospective
+
+```task
+id: CB-WP-0002-T07
+status: todo
+priority: low
+```
+
+Revise `specs/InnerLoop.md` and `specs/MetricsAndScenarios.md` from what
+this pass teaches about metric design. The specific question to answer:
+CB-WP-0001 produced a fully specified metric that could not be computed,
+and v1.0's "every metric names its instrument" rule was written to stop
+that. Did it? If a metric can still be written without a working
+instrument, the rule needs teeth — for example, requiring that the
+instrument be demonstrated on real data before the metric is accepted.