advise
Skill109 starsChanged 39 days agoMITView on GitHub
Solve problems using knowledge base insights - extracts search terms, runs parallel KB queries, synthesizes advice grounded in your own frameworks
3 other projects use a version of this.
Adopted by
Most recently taken up by
- yagizdo/quivertheir own version, first seen Nov 15, 2026
- felipeam86/pipe-skillstheir own version, first seen Nov 15, 2026
- hailneed/devpersonatheir own version, first seen Sep 4, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- felipeam86/pipe-skills★ 18same name
- yagizdo/quiver★ 16same name
- hailneed/devpersona★ 2same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository61 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- ai-crystallizeSkill
- analyze-kbSkill
- auto-discoverySkill
- benchmark-memorySkill
- brain-mergeSkill
- changelog-protocolSkill
- coherence-sweepSkill
- compute-lifecycleSkill
- create-articleSkill
- create-explanatory-imageSkill
- decideSkill
- deep-researchSkill
- detect-tensionsSkill
- dialecticSkill
- domain-watchSkill
- elicitation-techniquesSkill
- epistemic-classificationSkill
- epub-chapter-extractorSkill
- extract-document-insightsSkill
- extract-insightsSkill
- find-connectionsSkill
- first-runSkill
- get-perspective-onSkill
- get-youtube-transcriptSkill
- git-commit-pushSkill
- graduate-insightsSkill
- incubation-loopSkill
- ingest-sourceSkill
- insight-capture-formatSkill
- insight-interviewSkill
- integrate-recent-notesSkill
- learn-new-thingsSkill
- manage-reference-dataSkill
- manage-thinking-topicsSkill
- manage-watching-domainsSkill
- multi-format-book-extractorSkill
- nano-banana-image-generatorSkill
- propagate-changeSkill
- quick-searchSkill
- reasoning-checksSkill
- recallSkill
- ref-auditSkill
- ref-bridgeSkill
- ref-ingestSkill
- ref-querySkill
- ref-reconcileSkill
- ref-refreshSkill
- refresh-indexSkill
- ref-supersedeSkill
- scheduled-runSkill
- scope-mountSkill
- search-vaultSkill
- self-diagnosticSkill
- start-hereSkill
- synthesize-insightsSkill
- talkSkill
- test-memory-systemSkill
- think-about-itSkill
- update-changelogSkill
- The first 60 shown.
The file
We split this skill into the rules it gives an agent
8 rules. About 0 are common, 7 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈7
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 43
Search terms: decision making tradeoffs, explore exploit, focus prioritization, opportunity cost
- Distinctive: line 49
Run the searches in parallel (single message, multiple Bash calls):
- Distinctive: line 122
NO subagent spawning: - all work happens inline
- Distinctive: line 124
NO spreading activation: - use static search for speed
- Distinctive: line 125
Parallel execution: - run all searches in one message, all reads in the next
- Distinctive: line 127
Cite your sources: - always reference the specific notes used
- Distinctive: line 128
Be actionable: - don't just dump knowledge, apply it to the problem
- Distinctive: line 129
If KB lacks relevant content, say so honestly and offer general reasoning instead
≈ “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.
Files
Advise
Help solve problems by grounding advice in your accumulated knowledge and frameworks.
Purpose
Turn natural language problems into KB-grounded advice. Fast path: no subagents, no changelogs, no multi-layer expansion.
Routing: if the problem is a structured decision between courses of action ("X or Y?", go/no-go, "is it worth") → use /decide instead: it applies an explicit decision rule (ergodic filter, EV, robustness, value-of-information) and delivers tripwires, not just frameworks. /advise is for framing, understanding, and open problems.
Problem
$ARGUMENTS
Process
Step 1: Extract Search Terms (no tool calls - just reasoning)
From the problem description, identify 3-4 keyword clusters that would match relevant KB content:
- Core concepts (what domain is this?)
- Related frameworks (what mental models apply?)
- Analogous patterns (what similar problems exist?)
Example:
- Problem: "Should I focus on fundraising or product development?"
- Search terms:
decision making tradeoffs,explore exploit,focus prioritization,opportunity cost
Step 2: Parallel Knowledge Retrieval
Read role: reasoning (contract: scope-mount - never copy its tables, never run a bare search). Two passes per term in the same batch: core is the spine (the user's thinking), the reasoning mount is the evidence layer (what he has read). Before the batch, run the scope-mount trigger check on the problem text; a hit appends ,company / ,thinkers / ,Books/<slug> to the wide pass for this run only. Direction questions about the org are not yours - route to /canon-advise.
Run the searches in parallel (single message, multiple Bash calls):
BRAIN_READ_SCOPE=core resources/local-brain-search/run_search.sh "search term 1" --limit 3 --json
BRAIN_READ_SCOPE=core,Books,document-insights resources/local-brain-search/run_search.sh "search term 1" --limit 5 --json
BRAIN_READ_SCOPE=core resources/local-brain-search/run_search.sh "search term 2" --limit 3 --json
BRAIN_READ_SCOPE=core,Books,document-insights resources/local-brain-search/run_search.sh "search term 2" --limit 5 --json
BRAIN_READ_SCOPE=core resources/local-brain-search/run_search.sh "search term 3" --limit 3 --json
BRAIN_READ_SCOPE=core,Books,document-insights resources/local-brain-search/run_search.sh "search term 3" --limit 5 --json
resources/local-brain-search/run_connections.sh --hubs --json # fingerprint - always core; for Step 5's attractor check, same batchStep 3: Read Top Insights
From the search results, read 2-3 of the most relevant note files in parallel - when both passes returned, read at least one from each. A Books/ or Document Insights/ note is encountered material: cite it as what the user has read, never as his view (Check 3 in Step 5 enforces this):
# Use Read tool on the top-scoring, most relevant filesStep 3.5: Check BDG Context (optional, if top results are frameworks)
For any top result that looks like a framework or key insight, check its BDG context:
resources/brain-graph/run_brain_graph.sh inspect "Top Result Name" --jsonThis reveals: lifecycle phase (is it generative?), staleness (is it still fresh?), and typed edges (what does it drive?). Prioritize generative frameworks over reflective notes. Warn if citing a stale note.
Step 4: Synthesize Advice
Combine the retrieved insights to address the original problem:
- Apply frameworks from the notes to the specific situation
- Cite specific notes: [[Note Title]]
- Highlight tensions or tradeoffs the KB reveals
- Give concrete recommendations grounded in your own thinking
- Prioritize generative notes (lifecycle > 0.6) - these are the user's strongest frameworks
Step 5: Reasoning Checks (required)
Apply the shared contract in .claude/skills/reasoning-checks/SKILL.md before finalizing:
- Epistemic Inversion (always) - specific falsifier required; a generic hedge means redo it
- Attractor Check (always) - against the hubs fetched in Step 2; ≥2 top-10 hubs load-bearing → generate one non-attractor framing
- Provenance Base (always) - tally from the frontmatter of the notes read in Step 3; unendorsed synthesis must be labelled as such
- Reference Class (only if the advice hinges on a forecast, magnitude, or probability)
Append the resulting blocks after the Bottom line.
Output Format
## [Problem summary - one line]
**Relevant frameworks from your KB:**
- [[Note 1]] - [how it applies]
- [[Note 2]] - [how it applies]
- [[Note 3]] - [how it applies]
**My take (grounded in your insights):**
[2-4 paragraphs synthesizing advice, citing notes, applying frameworks to the specific problem]
**Key tradeoffs to consider:**
- [Tradeoff 1]
- [Tradeoff 2]
**Bottom line:** [One clear recommendation or framing]
[reasoning-checks blocks: Epistemic Inversion · Attractor Check · Provenance Base · Reference Class (when quantitative)]Rules
- NO subagent spawning - all work happens inline
- NO changelog creation - this is conversational, not archival
- NO spreading activation - use static search for speed
- Parallel execution - run all searches in one message, all reads in the next
- Maximum 2 rounds of tool calls - searches + hubs (parallel) + reads (parallel); the reasoning checks reuse those results, no extra round
- Cite your sources - always reference the specific notes used
- Be actionable - don't just dump knowledge, apply it to the problem
- If KB lacks relevant content, say so honestly and offer general reasoning instead
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
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.

