catalyst-respond

name	catalyst-respond
description	Process Catalyst detections — verify scope, format output, route to discussion or initiative creation. Internal skill — called when a system improvement opportunity is detected during conversation.
context	main
allowed-tools	["Read","Glob","Grep","Agent"]
spec-version	2

Catalyst Respond

Domain: initiative Version: 1.0.0

WHY

Catalyst detections need scope verification before surfacing to the user. Direct label output (💡 Catalyst:) without verification causes scope misattribution — improvement opportunities get tagged to the wrong squad or context. This skill is the mandatory gateway between detection and user-facing output.

Input

Required

• Detection signal: What was observed (friction, pattern, opportunity)
• Source context: Where the detection occurred (current task, squad, file path)

Available context

• workspace/squads/initiative/agents/catalyst.md — detection criteria and escalation thresholds
• workspace/squads/initiative/policies/catalyst-constraints.md — scope and frequency limits

Output

• Verified 💡 Catalyst: or 👤 Catalyst: label with correct scope attribution
• If initiative-scale: route to /initiative:start proposal
• If small: direct application proposal

STEPS

Step 1: Scope Verification

1. Identify which squad/domain the detection belongs to
2. Check if the improvement target actually exists (path verification)
3. Confirm the detection is not a duplicate of an existing initiative (workspace/initiatives/active/ scan)

Step 2: Scale Assessment

Indicator	Scale	Route
1-2 file edits, single squad	Small	Phase 2: Light Suggestion (inline)
3+ files, structural change	Medium	Phase 3: Discussion → Initiative proposal
Cross-squad, system-wide	Large	Phase 3: Discussion → Initiative proposal with subtopics
3+ commits on same topic (cumulative)	Escalation	Escalation suggestion

Step 3: Output

For improvement opportunities:

💡 Catalyst: [verified-scope] [one-line summary]
Evidence: [what was observed]
Scale: [Small/Medium/Large]

For user pattern observations:

👤 Catalyst: [one-line observation]
Suggested record: [target file in vault/user/]

For initiative-scale proposals:

💡 Catalyst: Current work has reached Initiative scale ([evidence]).
Propose: /initiative:start with topic "[suggested-topic]"?

JUDGMENT

When NOT to fire

• Current conversation is in the middle of focused execution — defer to session-end
• User has already rejected a similar suggestion this session
• Detection is about the Catalyst system itself (meta-observation loop prevention)

Plan Mode Interaction

• If plan mode is active: accumulate detections but do not output labels
• On ExitPlanMode: evaluate accumulated detections immediately
• If the approved plan is initiative-scale (5+ files, cross-project, structural): output escalation suggestion before execution begins

KNOWLEDGE REFS

Ref	Path	Purpose
Catalyst Agent	workspace/squads/initiative/agents/catalyst.md	Detection criteria, escalation thresholds
Catalyst Constraints	workspace/squads/initiative/policies/catalyst-constraints.md	Scope limits, frequency rules
Stage Transition	workspace/squads/initiative/policies/stage-transition-rules.md	Initiative creation prerequisites