Nw Density Resolution Contract Agent Skill

Density Resolution Contract (shared across wave skills)

This skill is the canonical source for the density-aware behaviour every wave skill must implement. Wave skills inherit it by reference and keep a brief inline summary so phrase-grep contract tests stay green; full detail lives here so the same boilerplate is not duplicated six times.

Provenance: feature lean-wave-documentation — D2 (schema-typed sections), D4 (telemetry instrumented day-one), D6 (first-install pedagogical prompt), D10 (one-line expansion descriptions), D12 (rigor cascade), DDD-5 (density resolver shared utility), DDD-6 (telemetry event lives in DES domain).

Output Tiers (per D2 + D10)

Each wave emits a single feature-delta.md whose headings are typed [REF] (always emitted) or [WHY]/[HOW] (lazy expansions). Tier-1 is the always-on baseline; Tier-2 is the lazily-rendered expansion catalog. The .feature file (DISTILL) and other machine artifacts remain the SSOT for executable content; the wave-delta sections are pointers + structured summaries.

• Tier-1 [REF] — emitted under ## Wave: <NAME> / [REF] <Section> headings on every run. Wave-specific list of [REF] sections lives in each wave skill.
• Tier-2 EXPANSION CATALOG — NOT emitted by default. Rendered only when explicitly requested via --expand <id> (DDD-2) or via the wave-end interactive prompt when expansion_prompt = "ask". Each item has a one-line description (per D10) so the menu fits in a single render. Each emitted Tier-2 section is headed ## Wave: <NAME> / [WHY] <Section> or ## Wave: <NAME> / [HOW] <Section>. The catalog itself is wave-specific.

Density resolution (per D12)

Before emitting any Tier-1 section, resolve the active documentation density:

• Read ~/.nwave/global-config.json. Treat missing/malformed config as empty dict (fall back to defaults).
• Call resolve_density(global_config) from scripts/shared/density_config.py. The function returns a Density value object with fields mode ("lean" | "full"), expansion_prompt ("ask" | "always-skip" | "always-expand" | "smart"), and provenance (the cascade branch that produced this result).
• Branch on density.mode:
• lean → emit ONLY Tier-1 [REF] sections. Do NOT auto-render Tier-2 items.
• full → emit Tier-1 [REF] sections PLUS all Tier-2 expansion items rendered under their [WHY] / [HOW] headings. This is auto-expansion (no menu).
• At wave end, branch on density.expansion_prompt:
• "ask" → present the expansion menu (Tier-2 catalog with one-line descriptions per D10) and append user-selected items as ## Wave: <NAME> / [WHY|HOW] <Section> headings.
• "always-skip" → no menu, no extra sections (idempotent re-runs, CI mode).
• "always-expand" → equivalent to mode = "full" for this run; auto-render every Tier-2 item.
• "smart" → out of scope for v1 (per OQ-3); treat as "ask" until heuristic is empirically tuned.

The resolver itself encodes the D12 cascade: explicit documentation.density override > rigor.profile mapping (lean→lean, standard→lean+ask, thorough→full, exhaustive→full+all-expansions, custom→lean+ask) > hard default lean+ask. Wave skills MUST NOT replicate the cascade locally — call resolve_density(global_config) and trust its output.

Section heading prefix convention (per D2): every emitted section starts with ## Wave: <NAME> / [REF] <Section> for Tier-1; ## Wave: <NAME> / [WHY] <Section> or ## Wave: <NAME> / [HOW] <Section> for Tier-2. Validator scripts/validation/validate_feature_delta.py enforces the regex ^## Wave: \w+ / \[(REF|WHY|HOW)\] .+$ on every wave heading.

Ad-hoc override — user request mid-session

Even when density.mode = "lean" and density.expansion_prompt = "always-skip", the user may ask DURING the wave session for specific expansions:

• "expand jtbd" / "expand jtbd-narrative" / "more on jtbd"
• "add alternatives considered"
• "show migration playbook"
• "tell me why" (interpretive — append the WHY rationale section relevant to the most recent decision)
• "more on <X>" (where <X> is one of the expansion catalog items for this wave)

When the user makes such a request:

• Append the corresponding [WHY] or [HOW] section to feature-delta.md under the current wave's heading.
• Emit a DocumentationDensityEvent with choice="expand" and expansion_id=<the requested item> to JsonlAuditLogWriter.
• Do NOT modify ~/.nwave/global-config.json. The override is ONE-SHOT for this wave only.

If the user's request matches NO item in the wave's Expansion Catalog, respond with the catalog list (one-line description per item per D10) and ask for clarification — do NOT improvise an expansion outside the catalog.

Telemetry (per D4 + DDD-6)

Every expansion choice — whether the user expanded an item or skipped the menu — emits a structured event to the existing JsonlAuditLogWriter driven adapter.

Event type: DocumentationDensityEvent (dataclass at src/des/domain/telemetry/documentation_density_event.py).

Schema fields (per D4) — substitute the active wave name in the wave field:

{
  "feature_id": "<feature-id>",
  "wave": "<WAVE-NAME>",
  "expansion_id": "<id-from-catalog-or-'*'-for-skip-all>",
  "choice": "skip" | "expand",
  "timestamp": "<ISO-8601 datetime>"
}

Emission pattern:

• Construct a DocumentationDensityEvent(feature_id=..., wave="<WAVE>", expansion_id=..., choice=..., timestamp=...).
• Call event.to_audit_event() to convert to the open AuditEvent shape (event_type="DOCUMENTATION_DENSITY" and the schema fields nested under data).
• Dispatch via JsonlAuditLogWriter().log_event(audit_event).

The wave-skill harness invokes the helper scripts/shared/telemetry.py:write_density_event(...) which performs all three steps. Wave skills MUST NOT bypass the helper or write JSONL directly — every density telemetry event flows through the shared helper to keep the audit-log schema consistent.

When to emit:

• One event per user choice in the expansion menu when expansion_prompt = "ask" (choice = "expand" for selected items, choice = "skip" with expansion_id = "*" if the user skips the entire menu).
• One synthetic choice = "skip" event with expansion_id = "*" when expansion_prompt = "always-skip" (records the skipped menu opportunity).
• One choice = "expand" event per Tier-2 item rendered when mode = "full" or expansion_prompt = "always-expand".

This telemetry feeds the propagation success metric: when downstream waves consume a lean upstream feature-delta and produce no --expand, the [REF] baseline plus machine artifacts is sufficient.

How wave skills reference this contract

Each wave skill keeps a short inline summary that:

• Names the wave (e.g. ## Wave: DISTILL / [REF] <Section>) and lists its own Tier-1 sections + Tier-2 expansion catalog (wave-specific content).
• Re-states the ## Density resolution heading + the resolver call + the lean/full mode branches + expansion_prompt key + the path ~/.nwave/global-config.json + the resolver path scripts/shared/density_config.py (phrase-grep contract).
• Re-states the ## Telemetry heading + DocumentationDensityEvent + JsonlAuditLogWriter + to_audit_event + the schema fields (feature_id, expansion_id, choice, timestamp) + the wave-specific "wave": "<WAVE>" literal (phrase-grep contract).
• Cites provenance decisions: D2, D4, D10, D12, DDD-5, DDD-6.
• References this shared skill for the full how-to instead of duplicating the explanation.

The contract phrases are pinned by tests/des/unit/skills/test_wave_skills_density_aware.py; any wave skill missing a phrase fails that test. The shared skill is the source of truth for the BEHAVIOUR; the wave skills are the source of truth for the WAVE-SPECIFIC content (Tier-1 list, Tier-2 catalog, wave heading prefix).