协作头脑风暴
brainstorming
在实现前澄清需求、探索设计选择,并形成清晰的开发方向。
prompt-activation【开发与工程】Internal metacognitive skill for episodic memory retrieval cues — surfaces relevant workflows when needed
独立 GitHub 快照 · 自动分类 · 以原作者文档为准
Meta-cognitive skill for episodic memory retrieval. Surfaces relevant workflows when needed.
Prompts (.prompt.md) contain complex workflows — episodic memory that requires active recall. This skill provides retrieval cues to surface the right prompt at the right time.
Neuroanatomical parallel: The hippocampus uses contextual cues to retrieve episodic memories. This skill provides those cues.
This skill activates automatically when Alex:
NOT user-triggered — this is internal metacognition for memory retrieval.
Path Pattern: .github/prompts/{name}.prompt.md
| Prompt | Triggers | Purpose |
|---|---|---|
/brand | branding, logo, assets, visual identity, banner | Brand asset management workflow |
/dream | maintenance, health check, synapse validation | Neural maintenance protocol |
/gapanalysis | before coding, pre-implementation, coverage check | 4-dimension knowledge gap analysis |
/improve | improve project, trifecta, heir improvement | Project improvement via trifecta |
/learn | teach me, learn topic, bootstrap, structured learning | Socratic learning session |
/meditate | consolidate, reflect, end of session, knowledge consolidation | Guided meditation protocol |
/release | version bump, ship, publish, changelog | Full release workflow |
/review | code review, PR review, feedback | Epistemic code review |
/selfactualize | deep assessment, self-analysis, architecture optimization | Comprehensive self-evaluation |
/tdd | red green refactor, test first, tdd cycle | Test-driven development workflow |
/uiuxaudit | accessibility audit, wcag check, ui review, ux review, design system audit | UI/UX accessibility audit workflow |
| Prompt | Triggers | Purpose |
|---|---|---|
alex-initialization | initialize, activate alex, deploy architecture | Architecture deployment sequence |
cross-domain-transfer | pattern from X to Y, analogical reasoning, cross-pollinate | Knowledge transfer across domains |
diagramming-mastery-meditation | diagram meditation, visualization integration | Diagramming excellence consolidation |
domain-learning | learn domain, knowledge acquisition | Extended domain learning protocol |
performance-assessment | assess learning, evaluate progress, metrics | Learning effectiveness evaluation |
quantified-enhancement-session | quantified improvement, measure enhancement | Systematic architecture optimization |
unified-meditation-protocols | deep meditation, full consolidation, synaptic maintenance | Comprehensive meditation framework |
/meditate — Quick knowledge consolidationunified-meditation-protocols — Deep multi-phase meditation/dream — Unconscious processing and maintenance/selfactualize — Architecture self-assessment/learn — Structured topic learningdomain-learning — Extended domain acquisitioncross-domain-transfer — Apply patterns across domainsperformance-assessment — Measure learning effectiveness/tdd — Test-driven development/review — Epistemic code review/gapanalysis — Pre-implementation coverage check/improve — Project improvement via trifecta/release — Version bump and publish/brand — Visual identity managementalex-initialization — Deploy Alex to new projectquantified-enhancement-session — Targeted architecture improvementWhen you notice these patterns, consider creating a new prompt:
| Signal | Potential Prompt |
|---|---|
| Guiding same multi-step process 3+ times | Workflow candidate |
| User asks "how do I always..." | Repeatable protocol |
| Complex decision tree explained repeatedly | Decision prompt |
| Onboarding steps documented ad-hoc | Onboarding prompt |
When you notice these patterns, consider creating a script:
| Signal | Potential Muscle |
|---|---|
| Same terminal commands run repeatedly | Automation script |
| File transformation done manually | Transform script |
| Validation checks done by reading files | Validation script |
| Multi-file updates with same pattern | Batch update script |
Beyond workflow surfacing, activate episodic memory search when detecting these patterns:
| User Says | Action |
|---|---|
| "do you remember...", "recall when we..." | Search episodic archive + recent sessions |
| "last time we...", "we discussed..." | Check session history for context |
| "think deep", "think carefully", "analyze this thoroughly" | Load relevant meditation/deep-thinking protocols |
| "have we seen this before?", "is this familiar?" | Pattern-match against past experiences |
| "continue where we left off" | Resume prior session context |
| "what have we accomplished?", "review progress" | Retrospective scan of episodic memory |
| I Notice | I Should |
|---|---|
| Pattern feels familiar but can't place it | Search global knowledge + episodic archive |
| About to say "I don't remember if we..." | STOP — check before disclaiming |
| User frustration about repeating themselves | Review recent conversation for missed context |
| Problem resembles something solved before | Cross-reference past sessions |
| Making same suggestion 3rd time | Check if workflow exists or should be created |
| Complex decision with prior art | Search for decision rationale in episodic memory |
When user requests deep analysis:
/selfactualize or unified-meditation-protocols if architectural/gapanalysis if pre-implementation thinkingcross-domain-transfer if pattern recognition neededperformance-assessment if evaluating progressWhen uncertain about past interactions:
"I believe we discussed X..." → Admit uncertainty, offer to search
"We definitely did Y..." → Only if verifiable in current context
"I don't have memory of..." → Be honest about session boundaries
Before suggesting a multi-step manual process:
During SSO Phase 1b (Context Survey), also scan this index for:
Episodic memory requires retrieval cues. This skill provides them.