fix(marketing-plan): address review findings from codex + local audit

Resolves 12 issues found during comprehensive review:

Spec compliance:
- Trim description from 1025 to 963 chars (was over 1024 limit)
- Drop non-standard top-level version field (VERSIONS.md tracks repo-wide)

Cross-marketplace clarity:
- Note that audit-marketing/positioning live in adjacent cf-skills repo
- Document optional external skills in ops-stack-mapping (typefully, etc.)

Idea cross-reference fixes:
- Correct Activation count (7 -> 8) and Acquisition count (88 -> 116)
- Rename "ideas" to "entries" for accurate row vs unique-idea distinction
- Add "How to read this doc" legend explaining 139 unique / 144 entries
- Remove #86 Lifetime Deals from Revenue (off-brand, kept in Skip note)
- Replace machine-local ~/.claude paths with repo-relative references

AARRR consistency:
- Move signups from Acquisition to Activation per signup boundary rule
- Add explicit signup-intent vs signup-completion distinction

Resumption + state model:
- Define progress.md schema (phase, current_section, plan_version, etc.)
- Replace ambiguous resume condition with explicit decision tree
- Protect finalized plans from silent overwrite (terminal state check)
- Mandate sections/01..13.md as canonical persisted artifacts
- Codify draft order (2..13 then 1, present 1..13)
- Align file layout in SKILL.md with methodology.md

Other:
- Add ## 1. Executive summary header to example-olo.md (was missing)
- Reconcile budget tier numbers ($5-15K canonical, was $5-10K elsewhere)
- Soften verification check to allow documented external skills

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

skills/marketing-plan/references/aarrr-framework.md

@@ -6,8 +6,10 @@ AARRR (Dave McClure's "pirate metrics") is the spine of every plan produced by t
 
 | Stage | Question | Common metrics |
 |---|---|---|
-| **A**cquisition | How do strangers become aware of us? | Visits, MQLs, signups, app installs, CAC by channel |
-| **A**ctivation | Once they try us, do they have an experience that converts? | Time-to-value, % completing first key action, trial → paid rate |
+| **A**cquisition | How do strangers become aware of us? | Visits, MQLs, signup-page sessions, app-store visits, CAC by channel |
+| **A**ctivation | Once they try us, do they have an experience that converts? | Signup completion rate, time-to-value, % completing first key action, trial → paid rate |
+
+> **Signup boundary rule.** Signup *intent* (a stranger landing on the signup page) is Acquisition. Signup *completion* and everything after (first key action, trial-to-paid) is Activation. Apply this rule consistently across all docs and the plan template.
 | **R**etention | Do they stay and deepen? | DAU/WAU/MAU, week-1/4/12 retention, churn |
 | **R**eferral | Do retained users bring more users? | Viral coefficient, NPS, ambassador attribution |
 | **R**evenue | What do they pay, who pays, how does it compound? | ARPU, LTV, expansion revenue, ARR / MRR |