fix council findings: vibe always --deep, debate in post-mortem, auto-escalation

Council validation surfaced 3 shared findings across both judges:
- /vibe now always uses --deep (3 judges) regardless of spec presence
- /post-mortem now documents --debate support for feature parity
- Added auto-escalation rule: preset perspective count auto-sets judge count
- Documented spec-compliance graceful degradation when no spec present

Author: boshu2

skills/council/SKILL.md

@@ -363,6 +363,14 @@ Your job is to find problems. A PASS with caveats is less valuable than a specif
 
 When `--preset` or `--perspectives` is used, judges receive the perspective-labeled prompt instead (see Agent Prompts section).
 
+### Auto-Escalation Rule
+
+**When `--preset` or `--perspectives` specifies more perspectives than the current judge count, automatically escalate judge count to match.** For example:
+- `/council --preset=security-audit validate X` → 3 perspectives → auto-escalate to 3 judges (equivalent to `--deep`)
+- `/council --perspectives="a,b,c,d" validate X` → 4 perspectives → auto-escalate to 4 judges (equivalent to `--count=4`)
+
+This prevents silently dropping perspectives. The `--count` flag overrides auto-escalation (user explicitly chose a count).
+
 ### Custom Perspectives
 
 Simple name-based:
@@ -420,6 +428,8 @@ code-review:
   error-paths:      "Trace every error handling path. What's uncaught? What fails silently?"
   api-surface:      "Review every public interface. Is the contract clear? Breaking changes?"
   spec-compliance:  "Compare implementation against the spec/bead. What's missing? What diverges?"
+  # Note: spec-compliance gracefully degrades to general correctness review when no spec
+  # is present in context.spec. The judge reviews code on its own merits.
 
 plan-review:
   missing-requirements: "What's not in the spec that should be? What questions haven't been asked?"

skills/post-mortem/SKILL.md

@@ -26,6 +26,7 @@ Two steps:
 /post-mortem --deep recent      # thorough council review
 /post-mortem --mixed epic-123   # cross-vendor (Claude + Codex)
 /post-mortem --explorers=2 epic-123  # deep investigation before judging
+/post-mortem --debate epic-123      # two-round adversarial review
 ```
 
 ---
@@ -86,10 +87,17 @@ The plan/spec content is injected into the council packet context so the `plan-c
 ```
 Single-agent structured review. Fast wrap-up without spawning.
 
+**With debate mode:**
+```
+/post-mortem --debate epic-123
+```
+Enables adversarial two-round review for post-implementation validation. Use for high-stakes shipped work where missed findings have production consequences. See `/council` docs for full --debate details.
+
 **Advanced options (passed through to council):**
 - `--mixed` — Cross-vendor (Claude + Codex) with retrospective perspectives
 - `--preset=<name>` — Override with different personas (e.g., `--preset=ops` for production readiness)
 - `--explorers=N` — Each judge spawns N explorers to investigate the implementation deeply before judging
+- `--debate` — Two-round adversarial review (judges critique each other's findings before final verdict)
 
 ### Step 4: Extract Learnings
 

skills/vibe/SKILL.md

@@ -121,11 +121,11 @@ If a spec is found, include it in the council packet's `context.spec` field:
 
 The spec content is injected into the council packet context so the `spec-compliance` judge can compare implementation against it.
 
-**Without spec — use default independent judges:**
+**Without spec — 3 independent judges (no perspectives):**
 ```
-/council validate <target>
+/council --deep validate <target>
 ```
-Falls back to council default behavior (independent judges, no perspectives).
+3 independent judges (no perspective labels). Vibe always uses `--deep` for consistency with /pre-mortem and /post-mortem.
 
 **Council receives:**
 - Files to review
@@ -139,11 +139,11 @@ Falls back to council default behavior (independent judges, no perspectives).
 ```
 Single-agent structured self-review. Fast, cheap, good for mid-implementation checks.
 
-**With --deep (3 judges, no spec):**
+**With explicit --deep (redundant — vibe always uses --deep):**
 ```
 /council --deep validate <target>
 ```
-3 independent judges (no perspective labels).
+3 independent judges (no perspective labels). Same as default vibe behavior.
 
 **With --mixed (cross-vendor):**
 ```
@@ -216,11 +216,12 @@ Each judge reviews for:
 
 | Judge | Verdict | Key Finding |
 |-------|---------|-------------|
-| Error-Paths | ... | ... (if spec found) |
-| API-Surface | ... | ... (if spec found) |
-| Spec-Compliance | ... | ... (if spec found) |
-| Judge 1 | ... | ... (if no spec — independent) |
-| Judge 2 | ... | ... (if no spec — independent) |
+| Error-Paths | ... | ... (with spec — code-review preset) |
+| API-Surface | ... | ... (with spec — code-review preset) |
+| Spec-Compliance | ... | ... (with spec — code-review preset) |
+| Judge 1 | ... | ... (no spec — 3 independent judges) |
+| Judge 2 | ... | ... (no spec — 3 independent judges) |
+| Judge 3 | ... | ... (no spec — 3 independent judges) |
 
 ## Shared Findings
 - ...