@thebassclef/lite 0.1.2 → 1.0.0

package/substrate/.claude/rules/plan-enumeration-needs-value-props.md

@@ -1,211 +0,0 @@
----
-tier: lite
-description: When the agent presents options the operator must pick from or evaluate, the whole proposal opens with a full pitch (claim + 3 reasons) and each option lands as a tweet (claim +…
----
-
-# Plan Enumeration Needs Value-Props
-
-When the agent presents options the operator must pick from or evaluate,
-the whole proposal opens with a full pitch (claim + 3 reasons) and each
-option lands as a tweet (claim + sharpener). **Bare names aren't options.**
-Names are labels; each label needs a value-prop attached.
-
-## Why this rule exists
-
-Operator scans on mobile + multi-pane desktop. Bare-name enumeration
-("Option A. WU-2", "Option B. #334", "Option C. WU-5") forces the
-operator to ask "what's WU-2?" — a follow-up turn that wouldn't be
-needed if the value-prop sat next to the name.
-
-Caught 2026-04-27e during /longrun prep — operator: "you need to land
-value props for each scope choice you propose going forward...we
-probably need to constrain/hook that in some way."
-
-The /value-prop skill already provides four compression modes (flash,
-tweet, brief, verbose). This rule wires them into how the agent
-presents enumerations, so plan output is scannable in one read.
-
-## When this rule fires
-
-Any agent output that enumerates options for the operator — "Pick A or
-B," "scope choices for this /longrun," "which approach do you want,"
-"three sizes:" — fires this rule.
-
-Specific surfaces:
-
-- `/longrun` (prep + checkpoint) — scope choices for the session
-- `/sprint` — next-work options
-- `/whats-the-plan` — chains of skills the agent intends to run
-- `/shape` — tier alternatives (quick / light / medium / full)
-- `/kiss` — bucket enumerations (MUST / SHOULD / COULD)
-- `/stage` — variant alternatives presented to operator
-- Any ad-hoc agent output where the operator picks among options
-
-Does NOT fire on:
-
-- Single-action replies (no options to pick from)
-- Tool-output relays (commit messages, test output)
-- Status reports without choice points (e.g., "shipped X")
-- Code content (diffs, file contents)
-
-## What the rule requires
-
-### The whole proposal opens with a full pitch
-
-Per `/value-prop` verbose mode: **claim + 3 reasons (why-it-matters)**.
-
-```markdown
-**The /longrun (full pitch):**
-
-Make bassclef's roadmap state machine-readable so /sprint and /longrun
-stop reconstructing scope from chronicles + git log.
-
-**Why it matters:**
-
-1. First-turn orientation drops from "read three chronicles + git log
-   + open PRs + guess" to one read.
-2. Two parallel sessions can't double-claim work — atomic claim/release
-   lands as a primitive.
-3. Operator stops manually clearing rescue files every session.
-```
-
-Three reasons, plain language, scannable.
-
-### Each option lands as a tweet
-
-Per `/value-prop` tweet mode: **claim + sharpener, ≤280 chars**.
-
-```markdown
-**Pick one:**
-
-- **A. WU-2 + WU-3 (~50t)** — Roadmap state becomes machine-readable.
-  Two parallel sessions can't pick the same work by accident. Claim/
-  release primitives ship.
-
-- **B. A + #334 (~70t)** — Plus a save-time check that catches agents
-  skipping the roadmap. Read-side and write-side determinism in one shot.
-
-- **C. Through WU-5 (~100t)** — Plus /sprint and /longrun actually use
-  the new reading. End-to-end this session.
-```
-
-Each option's value-prop sits next to its name. Operator reads, picks.
-
-### Bare names aren't options
-
-These are NOT acceptable enumeration shapes:
-
-```markdown
-Pick:
-- A
-- B
-- C
-```
-
-```markdown
-Pick:
-- WU-2
-- #334
-- WU-5
-```
-
-```markdown
-Three options: WU-A, WU-B, WU-C.
-```
-
-Each option needs a value-prop. If the value-prop is so obvious it
-doesn't need stating, the option probably isn't worth enumerating.
-
-## What the rule allows
-
-- **Plain language preferred** (composes with `/kiss words` — folds in what was the standalone `/grade10words` skill per bassclef#340).
-- **Tables when the options share a comparable axis** — e.g., "Tier / Estimate / Buys you" — but each row still needs the value-prop column populated.
-- **Skipping the full pitch when the options are presented in passing** — e.g., "Want me to ship A, B, or C?" mid-conversation. The full pitch is for proposal moments; the tweet-per-option discipline still applies.
-- **Inline alternatives** — "I'll ship X (or Y if you'd rather; Y trades speed for safety)" is fine; the alternative carries its sharpener inline.
-
-## Anti-patterns
-
-Direct quotes from sessions where this rule was missing:
-
-- "Pick one: A, B, or C" — three letters, no content
-- "WU-2 + WU-3 (~50t)" alone in a list — bare cost without value
-- "We could do this or that" without naming what each buys
-- "Three sizes — pick by how much determinism you want" without saying which size buys what
-
-## How the rule composes with existing skills
-
-- `/value-prop` is the primitive — flash/tweet/brief/verbose modes the
-  rule references directly.
-
-  INSTEAD of reinventing the format inline: call the skill. The four
-  compression modes already exist; reuse them.
-- `/kiss words` (bassclef#340 folded the planned `/grade10words` into `/kiss`) — substitutes plain language for jargon. Composes orthogonally; an option's tweet should already be in plain language by default.
-- `/feynman` — explanation skill. NOT a substitute for tweet-per-
-  option; /feynman is for "explain this concept," not "present these
-  choices."
-
-## Verification cue (self-check before output)
-
-Before presenting an enumeration, ask:
-
-1. Did I open with a claim + 3 reasons (full pitch)?
-2. Does each option carry a tweet (claim + sharpener)?
-3. Are any options just labels / IDs / bare names?
-4. Plain language? Or am I using "atomic" / "load-bearing" / "blast radius" without justifying?
-
-If any answer is wrong, rewrite before posting.
-
-## Eat-our-own-dogfood
-
-Every PR body that ships a bassclef-evolution change uses this format
-when it presents alternatives ("Out of scope," "Acceptance," "Test
-plan"). Each line earns its place; bare bullet enumerations get
-upgraded to tweets.
-
-Iteration-goal docs follow the same discipline in their "Steps
-enumerated" section.
-
-## Relationship to other rules
-
-- `feedback_lead_with_flash.md` (memory) — the operator-feedback
-  precedent; this rule codifies it as substrate
-- `visual-hierarchy.md` — long agent output uses tables/headers/HRs;
-  this rule adds the value-prop content discipline that makes the
-  hierarchy useful
-- `artifact-ingestion.md` — "Sources read" block discipline; this rule
-  is the analogous "options presented" discipline
-- `commit-conventions.md` — commit message format; this rule applies
-  to PR bodies, agent dialogue, and skill outputs (not commits — those
-  have their own format)
-- `sdlc-gates.md` — methodology-level gates; this rule fires before
-  any plan-presentation moment
-
-## Enforcement
-
-Methodology-level. The rule loads into every session via
-`additionalDirectories`. Agents self-check before posting an enumeration.
-
-If post-observation shows agents repeatedly presenting bare-name
-enumerations, a future hook could grep agent output for option-list
-shapes and BLOCK if value-prop content is missing — but the first line
-of defense is the rule loading into every session.
-
-## Override path
-
-`SKIP_PLAN_ENUMERATION_RULE=1 <command>` — logged to trace per existing
-bassclef hook idiom. Use only when the enumeration is genuinely
-stylistic / casual ("want coffee or tea?") and the value-prop discipline
-would be ceremony.
-
-INSTEAD of using the override for production decisions: write the
-value-prop. Production scope choices always carry weight worth
-articulating.
-
-## Retirement condition
-
-If observed across 6 months with zero operator corrections of
-enumeration formatting, this rule is correctly calibrated. If the
-operator stops asking "what does X buy me?" entirely, the rule has
-done its job.
-
-Closes bassclef#339.

package/substrate/.claude/rules/pr-body-shape.md

@@ -1,317 +0,0 @@
----
-tier: lite
-description: Every PR body opens with a ## Problem section drafted by /state-a-problem brief + --- + ## Goal.
----
-
-# PR Body Shape Contract
-
-Every PR body opens with a `## Problem` section drafted by `/state-a-problem brief` + `---` + `## Goal`. Detail comes AFTER. Operators and reviewers scan on mobile and multi-pane desktop. Without a problem-grounded opener, readers re-derive the reason for the PR every time. The skill is the contract. Hand-rolling reproduces the failure pattern bet 25k closed (see `docs/iteration-bets/2026-06-25k-discipline-visibility-and-pr-body-scrub-fix.md` § WU-1 L66 diagnosis).
-
-This rule closes two gaps:
-
-- **Mechanical.** Bassclef enforces what NOT to include in PR bodies (jargon BLOCK list + operator-private paths + strategic framing via `pr-body-scrub-check.sh`). No prior contract said what MUST appear at the opener.
-- **Methodology.** The prior Flash + Tweet pair (bassclef#906) was solution-framed twice — neither stated user pain. bassclef#1178 retires the pair at PR + issue surfaces and replaces both with a single `## Problem` section that centers user pain before solution.
-
-Sister to `.claude/rules/github-issue-flash-tweet.md` at the issue surface.
-
-## Why this rule exists
-
-`/diagnose` on 2026-06-10 surfaced two root causes for PR body drift:
-
-- **Root cause A** — no PR body shape contract. Authors improvised the first section name ("Goal", "What this PR ships", "What this PR delivers"). Sample drift across 12 recent PRs: 5 with explicit `## Goal`; 7 in other shapes.
-- **Root cause B** — the `/kiss` substitution table and the PR-body-scrub hook JARGON_TERMS list used different word sets. Closed by bassclef#907 (single source of truth in `standards/bassclef-internal-jargon.md`).
-
-V1 of this rule (bassclef#906) addressed Root cause A with Flash + Tweet + Goal. Operator observation on 2026-06-18 surfaced the gap. Flash and Tweet were both solution-framed. Neither stated user pain. Three UX research frameworks (NN/g, UX Design Institute, OpinionX) converge on a stable problem statement template (Who / What / When / Why-now / Outcome). bassclef#1178 retires Flash + Tweet at PR + issue surfaces and replaces both with a single `## Problem` section drafted by `/state-a-problem brief`.
-
-The shape contract still does what V1 did — gives readers a clean entry. The entry is now grounded in the problem, not in the solution claim.
-
-## Format contract
-
-Every PR body MUST open with:
-
-```markdown
-**Problem (≤500 chars):** <Who + What + When + Why now + Outcome — framework implicit; drafted by /state-a-problem brief>
-
----
-
-## Goal
-
-<What this PR ships, in one paragraph. Why now.>
-
-## Evidence
-
-- Source: <file path, ticket number, commit SHA — what backs the claim>
-- Warrant: <one sentence tying the source to the claim>
-```
-
-Format rules:
-
-- **Problem label is bold and enclosed in `**`** — a visual scan handle, not body prose
-- **Length limit enforced** — Problem ≤500 chars (matches `/state-a-problem brief` mode cap)
-- **Plain language** — no bassclef jargon in the Problem section. The Problem section is the FIRST place a reader meets the work. Jargon here destroys orientation. `/state-a-problem` runs `/kiss --grade10` inline by default so the rule is hard to violate
-- **Horizontal rule** (`---`) separates Problem from the structured body
-- **`## Goal` is the first section after the rule** — not `## Summary`, not `## What this PR ships`, not `## Why this matters`. Goal is the agreed label
-- **`## Evidence` follows Goal** — cites the source that backs the claim in Goal. Per `.claude/rules/assert-only-after-verify.md` — every operator-facing assertion carries source + warrant. Per `standards/opener-discipline.md` — the three-block opener extends to PR body via Problem + Goal + Evidence
-- **One paragraph per** — Problem is 3-4 sentences; Goal is one paragraph; Evidence is one bullet source + one bullet warrant
-
-## Required sections (in order)
-
-After Problem + `---`:
-
-| Section | Required? | Purpose |
-|---|---|---|
-| `## Goal` | **MUST** | One paragraph. What this PR ships. Why now |
-| `## Evidence` | **MUST** | Source + warrant per `.claude/rules/assert-only-after-verify.md`. Cites what backs the Goal claim |
-| `## Summary` | MUST when scope >1 file | Bullet list of substantive changes |
-| `## Test plan` | MUST | Checklist of how reviewer verifies |
-| `## Out of scope` | SHOULD | What this PR explicitly does NOT do |
-| `## /temperance + /luminary + /loop discipline` | **MUST under /longrun dispatch** | The three-skill cycle this PR ran (see below) |
-| `## Refs` | SHOULD | Issue numbers, related PRs, source chronicles |
-
-Other sections (Bundle justification, Inventory findings, Migration notes) appear between Goal and Test plan as needed. Drop sections that do not apply (e.g., Out of scope on a doc-only PR).
-
-## /temperance + /luminary + /loop discipline section (per bassclef-upstream#366)
-
-Every PR authored under `/longrun` dispatch MUST carry a section with the literal header `## /temperance + /luminary + /loop discipline`. The section makes the per-PR cycle named in `.claude/rules/loop-discipline.md` visible at PR review time.
-
-The header string is fixed and case-sensitive. The hook `pr-body-loop-discipline-check.sh` greps for it exactly. The section body names three things, one per bullet:
-
-- **/temperance** fired at `<surface>`. Scope decision answer in one line. Drift trigger in one line.
-- **/luminary** primary lens `<luminary-slug>` — one line on what the lens caught. Supporting lenses optional.
-- **/loop** iteration count `<N>`. Outcome — RED then GREEN, OR rework cycle if any.
-
-The fixed shape lets the grep check pass cleanly. It also gives readers a scannable three-line surface to verify discipline at PR review.
-
-Why this section exists: bets 25h + 25i + 25j ran the discipline floor. Markers landed under `state/markers/temperance/`, `state/markers/luminary/`, and `state/markers/loop/` per branch. But the PR bodies stayed silent. The operator caught the gap mid-goal 25j: "I'm not seeing /luminary listed for PR review and /loop iteration through to closure." The mechanical floor was correct. The visible bar was missing. This section closes the bar the moment a reader scans the PR.
-
-Forward-only enforcement. Already-merged PRs stay as is. Trivial-fix PRs under 3 files MAY skip the section per the Problem-section exception below.
-
-## What NOT to do
-
-- **Don't open with a section header** before Problem. INSTEAD: Problem is the headline. Section headers come after the `---` rule.
-- **Don't fall back to Flash + Tweet** — the V1 shape was solution-framed twice; neither stated user pain. INSTEAD: draft the Problem section via `/state-a-problem brief` mode. The skill produces the contract. The Who/What/When/Why-now/Outcome framework is implicit in the sentence shape it returns.
-- **Don't replace `## Goal` with synonyms.** INSTEAD: `## Goal` is the bassclef first section label after Problem. Reviewers and substrate-coherence checks grep for it.
-- **Don't bury the claim in `## Why this is needed` or `## Context` before Goal.** INSTEAD: the claim is the Problem + Goal pair. Context goes after.
-- **Don't use bassclef jargon in Problem or Goal.** INSTEAD: substitute plain English per `.claude/rules/plain-english-discipline.md` and `standards/bassclef-internal-jargon.md`. `/state-a-problem` runs `/kiss --grade10` inline by default. Hand-rolled drafts run `/kiss --grade10` separately before push.
-
-## Trivial-fix exception
-
-PRs with scope under 3 files AND no new SKILL.md AND no new rule MAY skip the Problem section. Trivial fixes (typo, env-var, off-by-one, one-line) often have no user-pain framing worth a section. The shape contract still requires `## Goal` as the opener in that case.
-
-The CI shape check skips the Problem requirement when the touched-file count is under 3 AND no SKILL.md or rule file is in the change.
-
-## Ticket-close keyword discipline (bassclef#974)
-
-When PR title carries `feat(#NNN):` or `fix(#NNN):` pattern, the body MUST include a `Closes #NNN` (or `Fixes #NNN` / `Resolves #NNN`) keyword. GitHub auto-closes tickets only when the closing keyword appears in the PR body — title alone does not trigger the close. Without the keyword, tickets stay open even after PRs merge, and the state-spine drifts from shipping reality.
-
-Per whereami#28b diagnose — of the last 20 merged PRs at 2026-07-29, only 4 carried a closing keyword in the body. 16 shipped work but left the ticket open (including #934 and #935 which required manual backfill during Session 28b).
-
-The `Refs` section is a good home for the keyword. Example:
-
-```markdown
-## Refs
-
-- Closes bassclef#123 (parent ticket — this PR ships the acceptance)
-- Refs bassclef#456 (related context — not closed by this PR)
-```
-
-Use `Refs` for context-only tickets that survive this PR. Use `Closes` / `Fixes` / `Resolves` for tickets whose acceptance criteria this PR satisfies fully.
-
-When a PR intentionally does NOT close its referenced ticket — for example, when the title cites a parent umbrella ticket but only ships one step — use `Refs #NNN` in the body and add a line explaining scope. Do NOT drop the closing keyword silently; document the choice.
-
-**Mechanical layer.** `pr-body-scrub-check.sh` Rule 7 fires as ADVISORY in V1. Reads `$PR_TITLE` env var passed by `pre-gh-pr-body-scrub.sh` wrapper. Extracts `feat|fix(#NNN)` pattern from title. Checks body for `(Closes|Fixes|Resolves)\s+#NNN`. Emits stderr note when missing. Exit 0 (does not block). Advisory in V1 while adopters get used to the discipline; may flip to blocking after calibration.
-
-## Vocabulary discipline at authoring time (bassclef#915)
-
-Per bassclef#915 V1: PR body authoring is a `/kiss` surface. Before invoking `gh pr create` (or `gh pr edit --body-file`), the agent runs `/kiss` words discipline against the body. It substitutes BLOCK and ADVISE-PHRASE terms from `standards/bassclef-internal-jargon.md`.
-
-- **BLOCK terms** — substitute per `/kiss --grade10` table; PR-body-scrub Rule 1 catches what slips
-- **ADVISE terms** — substitute per `/kiss --grade10` table; PR-body-scrub does not block on these
-- **ADVISE-PHRASE terms** — substitute or rephrase; PR-body-scrub Rule 5 emits advisory CI annotations on hits (does not block)
-- **ALLOW terms** (reserved per bassclef#763 — substrate, bassclef, cameo, Studio, Band, bandleader) — preserve
-
-V2 pre-push hook for `/kiss`-on-PR-body authoring ships per bassclef#1176 (PR #1179 merged 2026-06-18). Local PreToolUse Bash hook scrubs PR body before `gh pr create` reaches GitHub.
-
-## When this rule fires
-
-Every NEW PR body authored via:
-
-- `gh pr create` — local PreToolUse hook fires (bassclef#1176); CI hook fires as backstop (bassclef#739)
-- `gh pr edit` (body rewrite) — same shape contract applies
-- Manual operator authoring via GitHub web UI — `.github/pull_request_template.md` auto-fills the shape; operator amends
-
-Existing PRs are NOT retroactively backfilled. Forward-only enforcement keeps the discipline visible at the PR-creation surface where it lands.
-
-## Mechanical enforcement
-
-V1 shipped methodology + `.github/pull_request_template.md` template + the rule (bassclef#906). V2 shipped local + CI Bash-matcher hooks for jargon scrub (bassclef#1176 + bassclef#1177). V3 (this bassclef#1178 amendment) ships:
-
-- Shape contract enforces `## Problem` opener at feature-scope PRs
-- `pr-body-scrub-check.sh` extended with a shape check that flags a missing `## Problem` section on feature-scope PRs
-- `/state-a-problem brief` skill drafts the section (runs `/kiss --grade10` inline by default)
-- `.github/pull_request_template.md` template updated to show the Problem-section shape
-
-V4 (bassclef#974) ships Rule 7 — ticket-close keyword advisory. Local + CI hook layers extended:
-
-- `pre-gh-pr-body-scrub.sh` extracts PR title, passes as `$PR_TITLE` env var
-- `pr-body-scrub-check.sh` Rule 7 checks title-vs-body correlation for closing keyword
-- 4 new Tier 0 tests pin behavior (Tests 27-30 in `.claude/hooks/tests/pre-gh-pr-body-scrub.test.sh`)
-- Advisory in V1; V2 may flip to BLOCK after calibration
-
-## Override path
-
-`SKIP_PR_BODY_SHAPE=1 <command>` — for genuinely-not-PR-body cases (e.g., release-tag PRs, automated bot PRs, dependabot updates where the bot controls the body format). Logged via trace-helper.
-
-INSTEAD of overriding for tight scope: write the Problem section. Three sentences carrying the framework cost less than reviewer re-derivation.
-
-## Worked examples
-
-### Anti-pattern (Root cause A — what NOT to do)
-
-```markdown
-## Summary
-
-Add `/longrun` skill for autonomous long-running sessions...
-
-## Why this is needed
-
-Recurring pattern observed over March-April 2026 sessions...
-```
-
-Reviewer reads title → "Summary" header → 3 paragraphs to find the claim. Claim is buried. User pain stays unstated.
-
-### Anti-pattern (V1 Flash + Tweet — superseded by bassclef#1178)
-
-```markdown
-**Flash (≤180 chars):** Add `/longrun` skill — autonomous sessions paced by checkpoints and closeout discipline.
-
-**Tweet (≤280 chars):** Sessions over 50 turns drift without checkpoints. `/longrun` adds prep / checkpoint / closeout modes.
-```
-
-Better than no shape. But solution-framed twice. Reader still does not know WHY this work exists — only WHAT it does.
-
-### Conformant pattern (bassclef#1178 — Problem opener)
-
-```markdown
-**Problem (≤500 chars):** Today autonomous sessions over 50 turns drift without structured checkpoints — operators return mid-session to mid-stream context, no anchor on what shipped, what's blocked, or what's next. The /longrun skill adds prep / checkpoint / closeout modes that pace the session against named phase boundaries. The change we want is sessions over 50 turns that return clean state every time, not a context-archaeology task on every wake.
-
----
-
-## Goal
-
-Ship `/longrun` skill so autonomous sessions running >50 turns or across compaction boundaries have substrate-enforced pacing.
-
-## Summary
-
-- New `.claude/skills/longrun/SKILL.md` with three modes
-- ...
-```
-
-Reviewer gets user pain → solution → diff in 60 seconds. Detail follows for those who want depth.
-
-## Composes with
-
-- `.claude/skills/state-a-problem/SKILL.md` — drafts the Problem section in brief mode (≤500 chars); runs `/kiss --grade10` inline by default
-- `.claude/rules/github-issue-flash-tweet.md` (bassclef#484 — amended by bassclef#1178) — sister rule at issue-body surface
-- `.claude/rules/plain-english-discipline.md` (bassclef#790) — plain language requirement for Problem and Goal
-- `standards/pr-body-discipline.md` (bassclef#787) — scrub discipline at same surface
-- `.claude/rules/operator-facing-prose.md` (bassclef#790) — operator-facing prose discipline
-- `.claude/hooks/pr-body-scrub-check.sh` — sibling mechanical surface (CI hook); extended in bassclef#1178 with shape check
-- `.claude/hooks/pre-gh-pr-body-scrub.sh` — local pre-push hook (bassclef#1176)
-- `.github/pull_request_template.md` — auto-fill at PR-creation surface; updated for the Problem shape
-- bassclef#915 — phrase-shaped advisory perimeter (sister at vocabulary surface)
-- bassclef#907 — `/kiss` overhaul (single source of truth wordlist)
-
-## Distinct from
-
-- `commit-conventions.md` — commit messages have a separate format; this rule applies to PR BODIES, not commits
-- `iteration-bet-brief-completeness.md` — that's about delegation briefs (handoff content); this is about the PR-creation surface
-- `pattern-annotation.md` — that's source-code annotation discipline; orthogonal
-
-## Sources read
-
-- bassclef#906 — V1 parent ticket (Flash + Tweet + Goal shape; superseded at the opener)
-- bassclef#1178 — V3 amendment (Problem section + `/state-a-problem` skill)
-- bassclef#484 (`.claude/rules/github-issue-flash-tweet.md`) — sister discipline at issue surface
-- bassclef#787 (`standards/pr-body-discipline.md`) — sister discipline at scrub surface
-- bassclef#907 — `/kiss` overhaul (closes Root cause B)
-- bassclef#915 — phrase-shaped advisory perimeter
-- bassclef#1176 — V2 local PR-body scrub hook (PR #1179 merged 2026-06-18)
-- bassclef#1177 — V2 extends to gh issue + commit messages (PR #1180 merged 2026-06-18)
-- 2026-06-10 chronicle — V1 /diagnose finding
-- 2026-06-18 chronicle `chronicle/2026-06-18-v2-issue-and-commit-scrub.md` — V3 operator observation
-- NN/g, UX Design Institute, OpinionX problem-statement frameworks — the templates `/state-a-problem` synthesizes
-
-## Closes
-
-- V1 of bassclef#906 — Flash + Tweet + Goal shape
-- V2 of bassclef#906 — local + CI scrub at the opener (bassclef#1176 + bassclef#1177)
-- V3 of bassclef#906 — Problem section opener replacing Flash + Tweet (bassclef#1178 — this amendment)
-
-## Section-level DON'T / INSTEAD catalog (bassclef#1241)
-
-The catalog rows below encode authoring-time discipline per section. The wordlist scrub hook is the FLOOR. This catalog is the BAR. Apply the per-section catalog before the body ships rather than relying on the hook to catch what slipped past you.
-
-The catalog was drafted from two PR-authoring failures on 2026-06-19 (one wordlist-bypass via curl, one wordlist-as-bar drift). See bassclef#1241 for the source incidents.
-
-### Problem section
-
-Composes with `/state-a-problem brief`. Catalog rows below apply per-sentence inside the Problem block.
-
-| Don't | INSTEAD |
-|---|---|
-| Name script line numbers for internal mechanics ("L17 structural --execute gate") | Name the user-visible behavior ("the --execute flag is what makes it run") |
-| Third-person self-reference ("the operator can...", "the agent has no...") | Address the operator directly ("you can...", "I do not know your...") |
-| Bassclef-invented compounds (`discovery surface`, `registry shape`, `local-only discipline`, `Pattern-B'd`, `wiki-as-we-go discipline`, `source-of-truth`, `narrative moment`) | Plain restatement of what the compound names ("the screens the operator sees", "the file structure", "this rule that keeps chronicles local", "moved", "this rule that says check the wiki at every closeout", "the authoritative answer", "story worth telling") |
-| Buzz adjectives (`substantive`, `comprehensive`, `robust`, `leverage`) | Concrete data ("25-line block", "13/13 tests pass") |
-| Bassclef-internal nouns without gloss (`deferred-action`, `INFORM`, `Pattern B`) at first mention | First-mention gloss ("deferred-action file (the next session sees it as a TODO)") |
-| Nested parentheticals (`X (which is Y (per Z))`) | Split into separate sentences |
-
-### Goal section
-
-| Don't | INSTEAD |
-|---|---|
-| Methodology references the reader doesn't carry ("Phase 14 of the chain") | Name what ships and why now |
-| Multi-clause noun-piles ("substantive N-line block documenting A, B, C, D, and E") | Split into a bullet list |
-| Quote the rule's title as if it were the goal ("Apply bassclef#790 Rule 2 meta-discipline") | Name what the operator sees after merge ("/longrun's skill body reads in plain English") |
-
-### Summary section
-
-| Don't | INSTEAD |
-|---|---|
-| Invented compound verbs (`Pattern-B'd`, `scrubbed-and-shipped`) | Plain English ("moved to completed/", "rewrote and pushed") |
-| Code-fenced literals used as if prose (`` `_comment` replaced with substantive block ``) | Plain restatement ("Added a 25-line comment header that documents...") |
-| Restate the goal doc verbatim | Pick the 3-5 changes that matter to a reviewer |
-
-### Test plan section
-
-| Don't | INSTEAD |
-|---|---|
-| Assertions in bassclef-internal idiom ("Deferred-action surfaces at next session-start as INFORM") | Observable behavior ("Next session-start banner shows this under 'Info' (not 'Blocked')") |
-| Tense the reviewer can't verify without running ("hook fires clean") | What the reviewer can grep / run (`jq .pseudonyms[].pseudonym` returns `Falcon`) |
-
-### Out of scope section
-
-| Don't | INSTEAD |
-|---|---|
-| Third-person reference ("the agent cannot author...") | Direct ("I can't author entries — only you know your codenames") |
-| Defer-and-justify in the same line | Split: "X — not in this PR. Y — already tracked at #N" |
-
-### Refs section
-
-| Don't | INSTEAD |
-|---|---|
-| Long bassclef-internal qualifiers in parentheses | One-line ref per line, plain |
-
-### Section-cross-cutting rules
-
-| Don't | INSTEAD |
-|---|---|
-| Repeat literal filenames in title and body Summary together | Plain title; filenames in body where they're grep-anchored |
-| Mix `skillname` and `/skillname` for the same skill | Use `/skillname` consistently when referencing the skill; plain filename when referencing the file |
-| Use bassclef-internal verbs (`flip`, `Pattern-B'd`, `seed` as noun) | Plain verbs (`update`, `move`, `set up`) |
-| Drop session-invented shorthand (`Bucket 2a`, `Path B`, `S2`, `Option F`, `LR-1`) without an appositive on first mention per section | Gloss inline: `Bucket 2a (audit files with no tier tag at all)`; `S2 (Session 2 — Traceability Phase 1)`; per `.claude/rules/operator-facing-prose.md` Rule 1 extended per bassclef-upstream#1267 |
-
-## Retirement condition
-
-If observed across 6 months with zero operator-flagged shape drift on the Problem section, this rule is correctly calibrated. If `## Problem` first-section drift returns above 10% of sampled feature-scope PRs, the V3 CI shape check escalates from advisory to blocking. The catalog rows ship without an a-priori retirement condition — they encode discipline that survives any future skill or hook build.