agentleFS
JaimeParkerJaimeParker/stable-jarvis/skills/knowledge-distillation-from-discussion/SKILL.md

knowledge-distillation-from-discussion

Skill19 starsChanged 5 months agoMITView on GitHub

Distills a raw, text-based discussion (e.g., meeting notes, user-AI conversations in Markdown) into one or more structured, permanent 'evergreen' knowledge notes.

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 4 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Trigger: Use when the user wants to process a text-based discussion, conversation, or meeting notes to extract durable knowledge.

    line 12
  2. Distinctive:

    Follow this 4-stage process to execute the skill.

    line 21
  3. Distinctive:

    Confirm Source: Ask the user to provide the full path to the source discussion note within their Obsidian vault.

    line 25
  4. Very specific:

    Read Content: Use mcp_obsidian_read_note to load the content of the source file.

    line 27
  5. Distinctive:

    Structure Content: Create the Markdown content for the new note using the exact template below.

    line 49
  6. Very specific:

    Write Note: Use mcp_obsidian_write_note to save the new evergreen note to the confirmed path.

    line 74
  7. Very specific:

    Search: Use mcp_obsidian_search_notes with the query "MOC" to find potentially relevant Maps of Content.

    line 79
  8. Very specific:

    Patch: Use mcp_obsidian_patch_note or mcp_obsidian_read_note followed by mcp_obsidian_write_note to add the wikilink to the selected MOC(s).

    line 81
  9. Very specific:

    Tool: Use mcp_obsidian_move_note.

    line 84

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Skill: Knowledge Distillation from Discussion

Description

Distills a raw, text-based discussion (e.g., meeting notes, user-AI conversations in Markdown) into one or more structured, permanent 'evergreen' knowledge notes. This skill transforms a chronological 'fleeting note' into conceptual 'permanent notes', following the Zettelkasten/Evergreen methodology. It is designed to extract key insights, decisions, and concepts for a user's personal knowledge base (PKB).

Trigger: Use when the user wants to process a text-based discussion, conversation, or meeting notes to extract durable knowledge.

Input: The path to the source Markdown discussion note in the user's Obsidian vault. Output: Creates one or more new, atomic notes with declarative titles, updates relevant MOCs, and moves the original source note to the discussions archive.


Workflow

Follow this 4-stage process to execute the skill.

1. Isolate the Source (Fleeting Note)

  1. Confirm Source: Ask the user to provide the full path to the source discussion note within their Obsidian vault.
    • ask_user(question="Please provide the path to the Markdown discussion note you want to process.")
  2. Read Content: Use mcp_obsidian_read_note to load the content of the source file. Treat this content as a temporary "fleeting note".

2. Distill with AI-Powered Interrogation

  1. Select Distillation Lens: The goal is to extract specific types of insights, not just a generic summary. Ask the user what "lens" they want to apply to the discussion. Use ask_user with the following choices:
    • Core Concepts: "Find the fundamental principles, arguments, and counter-arguments."
    • Outcomes & Decisions: "Identify the central problem, the final decision, and any unresolved issues or action items."
    • Strategic Implications: "Analyze the discussion for strategic trade-offs, risks, and long-term benefits."
  2. Execute Distillation: Based on the user's choice, apply the corresponding advanced prompt to the text content you read in Step 1.

    • If "Core Concepts":

      "Analyze this discussion and extract the 1-3 core principles or 'atomic ideas' that were debated. For each idea, provide a one-sentence summary. Then, for each idea, detail the main supporting argument and any counter-arguments that were raised. Format the output so it can be structured into 'Core Insight' and 'Supporting Arguments' sections."

    • If "Outcomes & Decisions":

      "Review this discussion transcript. Format the output to clearly separate the 'Problem', 'Final Decision', and 'Action Items'. For the final decision, explain the consensus that was reached. For action items, list them as a Markdown checklist."

    • If "Strategic Implications":

      "Act as a CTO reviewing a technical discussion. Structure the output to clearly distinguish between 'Strategic Implications', 'Trade-offs', and 'Risks', based on the content of the discussion."

3. Synthesize New, Atomic Notes (Evergreen Notes)

  1. Propose Title: Based on the distilled output from the LLM, formulate a declarative title for a new, permanent note. The title should be a full sentence that captures the core insight (e.g., "Decoupling navigation and manipulation simplifies aerial grasping policies").
  2. Confirm Title and Path: Present the proposed title to the user and ask them to confirm or edit it. At the same time, ask for the desired folder path for this new permanent note (e.g., 30 Zettelkasten/32 Permanent/).
    • ask_user(questions=[...])
  3. Structure Content: Create the Markdown content for the new note using the exact template below. You MUST map the insights from the distillation lens chosen in Step 2 into the appropriate sections of this template.

    Template:

    ---
    tags: [distilled-insight, add-2-to-3-relevant-topic-tags]
    date: {{current_date}}
    type: evergreen
    ---
    # {{The Declarative Title}}
    
    **TL;DR:** [Write a one-sentence summary of the core insight or decision.]
    
    ### Context / Problem
    [Based on the distilled output, briefly describe the context or the problem that was being discussed.]
    
    ### Core Insight / Decision
    [Place the primary distilled knowledge here. For example, if the lens was 'Outcomes & Decisions', this section should contain the final consensus. If the lens was 'Core Concepts', this should contain the main principle.]
    
    ### Supporting Arguments / Caveats
    [Place secondary information here. For example, supporting arguments, counter-arguments, unresolved issues, or identified risks and trade-offs.]
    
    ---
    Source: [[50 Archive/Discussions/{{original_filename}}]]
  4. Write Note: Use mcp_obsidian_write_note to save the new evergreen note to the confirmed path.

4. Integrate and Archive

  1. Update MOCs:
    • Search: Use mcp_obsidian_search_notes with the query "MOC" to find potentially relevant Maps of Content.
    • Identify: Present the list of candidate MOCs to the user and ask which one(s) should be updated with a link to the new permanent note.
    • Patch: Use mcp_obsidian_patch_note or mcp_obsidian_read_note followed by mcp_obsidian_write_note to add the wikilink to the selected MOC(s). Provide a brief summary of the new note's context next to the link.
  2. Archive Original Note: Move the original source discussion note to the specific discussions archive.
    • Destination Path: 50 Archive/Discussions/ + (original filename).
    • Tool: Use mcp_obsidian_move_note. The oldPath is the source path from Step 1, and the newPath is the constructed archive path.
  3. Confirm Completion: Report back to the user, confirming that:
    • The new evergreen note has been created.
    • The relevant MOCs have been updated.
    • The original discussion note has been moved to 50 Archive/Discussions/.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.