competitor-discovery

Find new competitors worth adding to the comparison program.

Use this skill when the task is to discover, shortlist, or prioritize emerging products, not to fully research one specific competitor.

Load First

• Read comparison/TRACKING.md to avoid duplicate discovery work.
• Read comparison/COMPETITOR_GUIDE.md if you need the current relevance tiers or schema vocabulary.

What This Skill Does

• finds new competitors
• classifies them into the existing comparison landscape
• assigns an initial relevance tier
• recommends whether to add them to the tracker now, watch them, or ignore them

This skill does not produce a full canonical competitor profile. Hand off to competitor-research for that.

Discovery Workflow

1. Start from the current tracker and existing corpus.
2. Look for missing products in the most relevant clusters first:
   • agent orchestrators
   • first-party coding agents
   • desktop agent environments
   • PM tools that overlap with execution workflows
3. De-duplicate against:
   • comparison/<slug>/
   • comparison/_legacy/
   • aliases already noted in the tracker
4. For each candidate, decide:
   • category
   • what job it does
   • whether it is direct, adjacent, or monitor-only
5. Recommend one of:
   • add now
   • watch
   • ignore

Output Format

For each candidate, provide:

• name
• segment
• why it matters
• initial relevance tier: core, high, or monitor
• recommendation: add now, watch, or ignore

Keep discovery output concise. The point is triage, not full research.

Discovery Rules

• optimize for relevance, not exhaustiveness
• prefer active products over abandoned repos
• distinguish direct competitors from adjacent substitutes
• do not write a canonical profile here
• if a candidate looks important but facts are thin, still add it to the tracker with missing