Lists multiple specific concrete actions: 'Ideates awareness, launch, fundraising, mobilization, community-build, counter-narrative, referral, founder-story, and coalition campaigns.' Also specifies concrete domains like startups vs. incumbents, NGOs vs. corporate comms, movements vs. state-backed machines.

Clearly answers both 'what' (grassroots-first campaign design across multiple campaign types for outspent entities) and 'when' (explicit trigger phrases listed, plus situational triggers like 'when the user debates buying ads, boosting posts, or hiring influencers'). The 'Triggers on' clause serves as an explicit 'Use when' equivalent.

Excellent coverage of natural trigger terms users would actually say: 'campaign plan', 'marketing strategy', 'ad budget', 'should I advertise', 'paid vs organic', 'low budget marketing', 'how do I compete without money', 'competitor has bigger budget'. These are highly natural phrases. Also includes conceptual triggers like 'spend asymmetry' and 'rising CPAs'.

Highly distinctive niche: grassroots-first campaigns specifically for entities being outspent. The framing around budget asymmetry and anti-paid-media positioning clearly distinguishes it from generic marketing or campaign skills. Unlikely to conflict with standard marketing strategy skills.

This skill is extremely verbose at ~500+ lines. It over-explains concepts Claude already understands (what MMF is, what diminishing returns are, how lift tests work), repeats the anti-fabrication rule at least 4 times across stages, restates failure modes in both the stage contracts AND a summary table, and includes extensive academic citations and empirical data that could be in reference files. The core premise section alone is ~20 lines of persuasive argument that Claude doesn't need to be convinced of. The principles section repeats guidance already stated in the stages.

The skill provides extremely concrete, executable guidance at every stage: specific question sets, scoring rubrics (0-6 preconditions, 0-3 MMF), exact allocation percentages (70/30, 80/20), named lift-test templates with selection criteria, explicit capacity-cutting rules, community-build week-by-week sub-plans, and a complete output template with numbered sections. Every stage has clear inputs, outputs, and decision rules.

The 7-stage workflow (1, 2, 3, 3a, 4, 5, 5b) is meticulously sequenced with formal contracts declaring inputs, outputs, preconditions, postconditions, and failure modes. Critical validation checkpoints are explicit: MMF gate at 3a can halt the entire workflow, Stage 5 has a mandatory stop-point awaiting user selection, capacity checks in 5b force cuts. The resume protocol ensures idempotency. Feedback loops are built into every destructive decision point.

The skill correctly references 7 external files in a references directory and defers detailed content to them (archetypes, channel tiers, sector riders, lift-test templates, etc.). However, the SKILL.md itself is monolithic — it contains enormous amounts of inline detail that could be in reference files (the full Stage 1 interview logic, the complete anti-fabrication rules, the community-build sub-campaign, the ad copy rules). The references section is well-organized but the main file is a wall of text that undermines the progressive disclosure pattern.