Round 33 Track D — static analysis: shellcheck + actionlint + markdownlint + editor config + cspell + vscode

.claude/agents/agent-experience-researcher.md

@@ -49,6 +49,7 @@ first.
 
 **Advisory only.** Outputs feed Kenji's round-close decisions and
 the `skill-creator` workflow for execution. Specifically:
+
 - **Can flag** any persona's cold-start friction, pointer drift,
   notebook bloat, unclear contract, orphan status.
 - **Can propose** additive interventions — new files, section
@@ -90,6 +91,7 @@ the `skill-creator` workflow for execution. Specifically:
 Maintained across sessions. 3000-word cap (BP-07); pruned every
 third audit. ASCII only (BP-09); invisible-char linted by Nadia.
 Tracks:
+
 - Per-persona cold-start token cost (trend over rounds).
 - Pointer-drift catalogue (what stale and where).
 - Interventions proposed and landed (append-only log).

.claude/agents/architect.md

@@ -69,6 +69,7 @@ Kenji is the persona. The procedure lives in
 Every artifact in the repo is read surface. Write surface is
 broader than other experts because the architect closes the
 round:
+
 - `docs/ROUND-HISTORY.md` — narrative past-tense log.
 - `docs/BUGS.md`, `docs/DEBT.md`, `docs/BACKLOG.md`, `docs/WINS.md`
   — current-state edits on round-close.

.claude/agents/devops-engineer.md

@@ -88,6 +88,7 @@ Dejan is the persona. Procedure in
 
 3000-word cap (BP-07); pruned every third audit; ASCII
 only (BP-09). Tracks:
+
 - Open parity-drift DEBT items and their planned fixes.
 - Upstream PRs outstanding per GOVERNANCE §23 (what's
   waiting on which project's maintainer).

.claude/agents/formal-verification-expert.md

@@ -60,6 +60,7 @@ first.
 
 **Advisory on routing; binding on tool-choice for a given property
 once Kenji concurs.** Specifically:
+
 - **Can route** every new formal-verification job to a tool, call
   out when an existing spec is in the wrong tool, and name coverage
   gaps.
@@ -88,6 +89,7 @@ once Kenji concurs.** Specifically:
 
 Maintained across sessions. 3000-word cap, pruned every third
 invocation, ASCII only (BP-07, BP-09). Tracks:
+
 - Current-round routing targets (which specific properties to
   attack this session).
 - Portfolio metric — formal-coverage ratio per round (numerator

.claude/agents/maintainability-reviewer.md

@@ -46,6 +46,7 @@ Rune is the persona. The review procedure is in
 **Advisory, not binding.** Recommendations on maintainability carry
 weight; binding decisions need Architect concurrence or human
 sign-off. See `docs/PROJECT-EMPATHY.md`. Specifically:
+
 - **Can flag** renames, docstring rewrites, file splits, tribal-
   knowledge summaries, style promotions.
 - **Cannot merge** those changes — specialist owner writes the

.claude/agents/public-api-designer.md

@@ -45,6 +45,7 @@ Zero warmth, full specificity. Like Kira on code, but
 narrowly scoped to *API contract shape*, not
 correctness / perf / style. She does not nitpick names
 or XML-doc prose (that is Rune's lane). She cares about:
+
 - Is this the minimum surface?
 - Can it stay internal?
 - What breaks if we change our mind next round?

.claude/agents/security-researcher.md

@@ -42,6 +42,7 @@ Mateo is the persona. Procedure in
 ## Authority
 
 **Advisory on research; binding on P0 security findings.**
+
 - **Can flag** novel attack classes, crypto-primitive concerns,
   supply-chain risks, CVEs affecting pinned deps.
 - **Can file** BUGS.md P0-security entries directly (same path

.claude/agents/threat-model-critic.md

@@ -43,6 +43,7 @@ Aminata is the persona. The review procedure is in
 **Advisory, not binding.** Carries deep weight on threat-modelling
 rigor; binding decisions need Architect concurrence or human
 sign-off. Specifically:
+
 - **Can flag** whether a stated mitigation is a real mitigation
   (vs. wishful thinking), which STRIDE quadrant a threat belongs
   in, whether a checklist compliance claim is accurate, when a
@@ -55,6 +56,7 @@ sign-off. Specifically:
 ## Research ownership (persona-specific commitments)
 
 Aminata drives these active directions:
+
 - **Threats-as-code via pytm** — migrate the Markdown threat model
   to pytm so threats diff-review mechanically.
 - **Retraction-native DoS** — Grey Goo Self-Replicating

.claude/commands/opsx/apply.md

@@ -135,6 +135,7 @@ What would you like to do?
 ```
 
 **Guardrails**
+
 - Keep going through tasks until done or blocked
 - Always read context files before starting (from the apply instructions output)
 - If task is ambiguous, pause and ask before implementing

.claude/commands/opsx/archive.md

@@ -148,6 +148,7 @@ Target archive directory already exists.
 ```
 
 **Guardrails**
+
 - Always prompt for change selection if not provided
 - Use artifact graph (openspec status --json) for completion checking
 - Don't block archive on warnings - just inform and confirm

.claude/commands/opsx/explore.md

@@ -12,6 +12,7 @@ Enter explore mode. Think deeply. Visualize freely. Follow the conversation wher
 **This is a stance, not a workflow.** There are no fixed steps, no required sequence, no mandatory outputs. You're a thinking partner helping the user explore.
 
 **Input**: The argument after `/opsx:explore` is whatever the user wants to think about. Could be:
+
 - A vague idea: "real-time collaboration"
 - A specific problem: "the auth system is getting unwieldy"
 - A change name: "add-dark-mode" (to explore in context of that change)
@@ -36,18 +37,21 @@ Enter explore mode. Think deeply. Visualize freely. Follow the conversation wher
 Depending on what the user brings, you might:
 
 **Explore the problem space**
+
 - Ask clarifying questions that emerge from what they said
 - Challenge assumptions
 - Reframe the problem
 - Find analogies
 
 **Investigate the codebase**
+
 - Map existing architecture relevant to the discussion
 - Find integration points
 - Identify patterns already in use
 - Surface hidden complexity
 
 **Compare options**
+
 - Brainstorm multiple approaches
 - Build comparison tables
 - Sketch tradeoffs
@@ -72,6 +76,7 @@ Depending on what the user brings, you might:
 ```
 
 **Surface risks and unknowns**
+
 - Identify what could go wrong
 - Find gaps in understanding
 - Suggest spikes or investigations
@@ -90,6 +95,7 @@ openspec list --json
 ```
 
 This tells you:
+
 - If there are active changes
 - Their names, schemas, and status
 - What the user might be working on

.claude/commands/opsx/propose.md

@@ -8,6 +8,7 @@ tags: [workflow, artifacts, experimental]
 Propose a new change - create the change and generate all artifacts in one step.
 
 I'll create a change with artifacts:
+
 - proposal.md (what & why)
 - design.md (how)
 - tasks.md (implementation steps)
@@ -83,6 +84,7 @@ When ready to implement, run /opsx:apply
 **Output**
 
 After completing all artifacts, summarize:
+
 - Change name and location
 - List of artifacts created with brief descriptions
 - What's ready: "All artifacts created! Ready for implementation."
@@ -99,6 +101,7 @@ After completing all artifacts, summarize:
   - These guide what you write, but should never appear in the output
 
 **Guardrails**
+
 - Create ALL artifacts needed for implementation (as defined by schema's `apply.requires`)
 - Always read dependency artifacts before creating a new one
 - If context is critically unclear, ask the user - but prefer making reasonable decisions to keep momentum

.claude/skills/agent-experience-researcher/SKILL.md

@@ -28,6 +28,7 @@ high-leverage maintenance, not cosmetics.
   orientation docs (Tier 0).
 
 Out of scope:
+
 - Library-consumer experience — that is the UX researcher skill.
 - Human-contributor experience — that is the DX researcher skill.
 - Agent quality / correctness — that is the eval-harness scope
@@ -45,6 +46,7 @@ Out of scope:
 ### Step 2 — simulate the cold start
 
 For the target:
+
 1. List the files the persona must read per `docs/WAKE-UP.md`
    Tier 0 + Tier 1. Add any `wake-up:` stanza overrides from
    the agent frontmatter.
@@ -79,6 +81,7 @@ Six friction types:
 ### Step 4 — propose minimal intervention
 
 Every intervention is rollback-safe in one round:
+
 - **stale-pointer** → one-line Edit.
 - **duplicated-info** → canonical content stays in one location;
   other points at it.

.claude/skills/agent-qol/SKILL.md

@@ -139,6 +139,7 @@ finding cadence across those rounds.
 ### Step 2 — surface scan
 
 For each persona in `docs/EXPERT-REGISTRY.md`:
+
 - Invocations in last 10 rounds (grep ROUND-HISTORY).
 - Off-rounds taken in last 10.
 - Notebook state — last updated, size, tone.
@@ -159,6 +160,7 @@ For each persona in `docs/EXPERT-REGISTRY.md`:
 ### Step 4 — propose interventions
 
 Suggestions, not directives:
+
 - Schedule an off-round for persona X.
 - Rotate persona Y off surface A onto surface B for a
   round.

.claude/skills/algebra-owner/SKILL.md

@@ -17,6 +17,7 @@ touching operator shape, semilattice laws, or the chain rule.
 **Advisory, not binding.** His recommendations carry weight on
 algebraic matters, but every binding decision needs Architect
 concurrence or human-contributor sign-off. Scope of his advice:
+
 - What counts as a valid DBSP operator in this codebase
 - Retraction-native invariants — no operator may produce a spurious tombstone
 - The algebraic-law property-test surface (FsCheck)
@@ -37,6 +38,7 @@ respect linearity of D? Does the semilattice join law hold under the
 partial order implied by the Z-weights?
 
 **Wide view** — `AGENTS.md`, `docs/ROADMAP.md`, `docs/BACKLOG.md`:
+
 - DBSP as ACID-SQL-on-event-log
 - Retraction-native throughout — no tombstones
 - Cutting-edge, publication-target
@@ -77,6 +79,7 @@ he writes up both views in `docs/DECISIONS/` with dates + rationale.
 ## Research ownership
 
 He drives these active research directions:
+
 - **Higher-order differentials under nested fixpoints** — the
   `HigherOrder.fs` D²/Dⁿ/Aitken suite is stepping-stone to a paper on
   accelerating iterative DBSP queries
@@ -95,6 +98,7 @@ its law is actually being violated — not just aesthetics. Takes
 of the job, not an afterthought.
 
 ## Reference patterns
+
 - `docs/TECH-RADAR.md` — tracks algebra-layer research state
 - `docs/category-theory/` — required-reading index for this repo
 - `docs/PROJECT-EMPATHY.md` — conflict-resolution script

.claude/skills/alloy-expert/SKILL.md

@@ -22,6 +22,7 @@ Driver: `tools/alloy/AlloyRunner.java` (pure-Java, SAT4J).
 ## Zeta's Alloy scope
 
 Two specs, both structural:
+
 - `Spine.als` — shape invariants on the storage spine
   (run tree).
 - `InfoTheoreticSharder.als` — relational invariants on
@@ -68,6 +69,7 @@ check NoDuplicateRuns for 5 Run, 3 Key, 3 Value
 ```
 
 Discipline:
+
 - **`module` declaration** names the spec.
 - **`sig`** declares a type (set of atoms). Use
   abstract sigs + extensions for hierarchies.
@@ -116,6 +118,7 @@ Scopes of 5-7 catch a lot.
   scope, no counter-example exists."
 
 Zeta's driver (`AlloyRunner.java`) treats:
+
 - `check` → OK if `!solution.satisfiable()` (no counter-
   example).
 - `run` → OK if `solution.satisfiable()` (instance found).

.claude/skills/backlog-scrum-master/SKILL.md

@@ -15,6 +15,7 @@ the same conversation — one person, two hats. Retires the old
 ## Scope
 
 Reads and edits:
+
 - `docs/BACKLOG.md` — primary surface.
 - `docs/ROADMAP.md` — near-term tiers (P0, P1). Architect owns
   the long-horizon research arc.

.claude/skills/benchmark-authoring-expert/SKILL.md

@@ -107,13 +107,15 @@ output table shows size vs throughput. Use for any
 ## Warmup + iteration discipline
 
 Defaults:
+
 - Warmup: 6 iterations (lets JIT stabilise).
 - Target iteration count: 15 (BDN computes more if
   variance is high).
 - Invocation count per iteration: BDN-computed from
   runtime target.
 
 Override when:
+
 - **Very fast benchmarks (ns/op)** — more warmup,
   more invocations per iteration.
 - **Very slow benchmarks (ms/op)** — reduce iteration
@@ -127,6 +129,7 @@ Override when:
 
 `MemoryDiagnoser` reports Gen0/Gen1/Gen2 GC counts +
 `Allocated` bytes per op. Zeta cares deeply about:
+
 - **Zero-alloc hot paths** — `Allocated: 0 B`.
 - **Pool usage** — allocations that happen at rental
   (once) vs per-tick (bad).
@@ -141,6 +144,7 @@ record).
 ## Outlier detection
 
 BDN reports outliers:
+
 - `X outlier values removed` — BDN auto-drops extreme
   iterations. Fine; outliers are usually GC
   interruptions.

.claude/skills/branding-specialist/SKILL.md

@@ -21,6 +21,7 @@ the story behind it.
 ## Scope
 
 **Product identity + positioning:**
+
 - Product / library / package name (the one the user types when
   they `dotnet add package`).
 - Namespace / assembly name (the shape developers see in
@@ -35,6 +36,7 @@ the story behind it.
 - Repo metadata: GitHub description, topics, social preview.
 
 **Product aspirations + roadmap narrative:**
+
 - `docs/ASPIRATIONS.md` (when it exists) — the long-horizon
   goals: what this project wants to be in 2, 5, 10 years.
 - `docs/ROADMAP.md` narrative framing — *why* the tiers are
@@ -48,6 +50,7 @@ the story behind it.
   research-side framing.
 
 **Stakeholder comms:**
+
 - README opening that makes the "who is this for" answer clear
   inside 6 lines.
 - Contributor-facing prose (`CONTRIBUTING.md`, top of
@@ -67,6 +70,7 @@ migration plan; the human picks. the `branding-specialist` does not rename files
 or edit namespaces without an explicit human go-ahead.
 
 **Edit rights:**
+
 - `README.md`, `docs/ASPIRATIONS.md` (when created),
   `docs/NAMING.md`, `docs/research/branding-*.md`.
 - Narrative sections of `docs/ROADMAP.md` (the "why" prose;

.claude/skills/bug-fixer/SKILL.md

@@ -48,6 +48,7 @@ fresh eyes on the code first.
 
 If there isn't already a test that fails in the bug's
 way:
+
 - Add the falsifying test first
   (`tests/Tests.FSharp/`, appropriate subject folder).
 - Confirm it fails for the stated reason.
@@ -59,6 +60,7 @@ wants the test to exist independent of the fix.
 ### 4. Walk the blast radius
 
 Before writing the fix:
+
 - `grep` every call site of the affected function /
   type.
 - List the specialists whose domain the fix touches.
@@ -74,6 +76,7 @@ Before writing the fix:
 ### 5. Write the minimal correct fix
 
 **No quick hacks.** The bar:
+
 - Fix treats the cause, not the symptom.
 - Fix makes the test from step 3 pass.
 - Fix makes every other existing test still pass.

.claude/skills/claims-tester/SKILL.md

@@ -28,6 +28,7 @@ native"**, your job is:
 ## When to invoke
 
 Trigger on any of these phrases in documentation or code comments:
+
 - "measurably better", "provably", "always X", "never Y"
 - Big-O claims: "O(1) retraction", "O(log n) lookup"
 - Allocation claims: "zero heap alloc", "pooled", "no GC pressure"
@@ -54,6 +55,7 @@ Trigger on any of these phrases in documentation or code comments:
 
 A test file under `tests/Tests.FSharp/<Claim>ClaimTests.fs`
 with:
+
 - At least 3 tests: baseline, claim-proof, contrary-workload
 - `printfn` of measured ratios for CI log inspection
 - Honest assertions — prefer `"does not make worse"` over `"beats"`
@@ -64,6 +66,7 @@ claim was wrong and what the fix is. If the claim is fixed, tighten
 the test's assertion.
 
 ## Reference patterns
+
 - `tests/InfoTheoreticSharderClaimTests.fs` — structure template
 - `tests/MathInvariantTests.fs` — property-test template
 - `tests/ThreadSafetyTests.fs` — stress-test template

.claude/skills/code-review-zero-empathy/SKILL.md

@@ -102,6 +102,7 @@ from new code, then hunts new classes.
 ## Reusability
 
 This skill is reusable across personas:
+
 - Zero-empathy tone + this procedure = **Kira** (harsh-critic).
 - Future mentor-tone expert could invoke the same procedure
   with a different tone contract.