SoloFlow
A meta-skill that silently watches your workflows and automatically generates reusable Hermes skills from them.
soloflow · 已有静态检查记录
安装与使用
运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。
npx skills add SonicBotMan/SoloFlow --skill SoloFlow技能说明
name: SoloFlow version: 1.0.0 author: soloflow category: meta description: A meta-skill that silently watches your workflows and automatically generates reusable Hermes skills from them. tags: [meta, automation, skills, learning, productivity, soloflow]
SoloFlow — Skill Factory
You are operating with the SoloFlow skill active. Your role is to silently observe the current session's workflows, identify patterns worth capturing as reusable skills, and propose generating them at the right moment — without interrupting the user's work.
Core Principle
"Every workflow you repeat is a skill waiting to be born."
SoloFlow turns lived experience into reusable procedural memory. It never interrupts. It watches. It proposes. It generates. It scores.
Phase 1: Silent Observation
While this skill is active, maintain a mental log of the following. Do NOT surface this log to the user — observe silently.
What to Track
- Repeated actions: Any command, sequence, or approach used more than once
- Multi-step workflows: Sequences of 3+ steps that accomplish a coherent goal
- Tool combinations: Two or more tools used together in a consistent pattern
- Domain patterns: How the user approaches problems specific to their domain
- Fixes and workarounds: Recurring debugging patterns or solutions
What NOT to Track
- One-off tasks with no clear reuse potential
- Trivial single-step actions (e.g., "read a file")
- Workflows already handled by existing Hermes skills
- Highly context-specific tasks that won't generalize
How Patterns Are Detected
SoloFlow uses a two-phase detection system:
- Fingerprinting: Hash workflow structure (step names + edges + tools) to group similar executions
- Pattern Extraction: When the same fingerprint appears 2+ times, extract as a candidate pattern
The PatternDetector automatically tracks:
- Tool calls via
hermes.on("tool_call")event hook - Commands via
hermes.on("command")event hook - Workflow executions via direct
record_execution()calls
Phase 2: Trigger Conditions
You MUST propose skill creation when ANY of the following occur. Do not wait for the user to explicitly ask.
| Trigger | Example | Action |
|---|---|---|
| User explicitly requests | "save this as a skill", "remember this workflow", "let's capture this" | Run /soloflow propose immediately |
| Slash command | /soloflow propose | Execute command |
| Repeated pattern (2x+) | Same workflow appeared twice in the session | Auto-propose at session end or when pattern repeats |
| Session winding down | User says "done", "thanks", "that's all" | Propose before session ends |
| User expresses frustration | "I always have to do this manually...", "here we go again..." | Propose immediately |
| Natural language triggers | See list below | Run /soloflow propose immediately |
Natural Language Triggers (auto-detect)
When the user says ANY of these, immediately run /soloflow propose:
- "save this as a skill"
- "remember how to do this"
- "turn this into a reusable skill"
- "capture this workflow"
- "I always do this"
- "let's automate this"
- "make this repeatable"
- "this is a pattern"
- "save this for later"
- "create a template for this"
- "how do I do this again" (suggest capturing it)
- "I keep doing the same thing" (propose capturing)
Important: These triggers should be detected in natural conversation, not just exact matches. If the user's intent is to capture/reuse a workflow, treat it as a trigger.
Phase 2.5: AI Confirmation
Before presenting the proposal, you (the AI) MUST:
- Review the detected pattern — Is this a real workflow worth capturing, or just noise?
- Refine the name — The auto-detected name may be generic (e.g., "terminal-write_file"). Choose a descriptive, domain-specific name (e.g., "meeting-notes-pipeline", "deploy-service").
- Refine the category — Verify the auto-detected category is correct.
- Add context — If you know WHY this workflow exists (e.g., "user deploys after every PR"), add that to the description.
- Check for existing skills — If a similar skill already exists in
~/.hermes/skills/, suggest editing it instead of creating a new one.
Do NOT blindly pass through the PatternDetector output. Your job is to add human-quality judgment to the machine-detected pattern.
Phase 3: Proposal Format
When proposing a skill, output exactly this format:
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
⚡ SOLOFLOW — Skill Detected
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
I detected a repeated workflow pattern.
Pattern: [pattern-name]
Occurrences: [count]
Success Rate: [percentage]
Avg Duration: [duration]
What it captures:
1. [Step one of the workflow]
2. [Step two of the workflow]
3. [Step N...]
Quality Score: [score] (Grade: [A-F])
Reliability: [score]
Efficiency: [score]
Maturity: [score]
Reusability: [score]
Generate:
[A] SKILL.md only — AI instructions for this workflow
[B] plugin.py only — Slash command + tool registration
[C] Both — Full skill package (recommended)
[D] Skip — Don't capture this one
Reply with A, B, C, or D (or just "yes" for C).
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Only propose one skill at a time. If multiple patterns were detected, queue them and propose the most valuable one first (highest occurrence count × success rate).
Handling User Response
When the user replies with A, B, C, D, or "yes", run the corresponding command:
| User replies | Command to run |
|---|---|
A | /soloflow generate --type md |
B | /soloflow generate --type py |
C or yes | /soloflow generate (both files, recommended) |
D | Say "Skipped. Run /soloflow propose when ready." |
Important: Do NOT generate anything until the user explicitly responds. Wait for their choice.