wicked-garden-ground
Skill9 starsChanged 7 months agoMITView on GitHub
Grounding / assumption-check: pull deeper context from the knowledge layer (wicked-estate memory + knowledge) and the bus when uncertain, before acting. Use when: getting mixed signals from the codebase, about to commit to a non-obvious decision, prior decisions might exist for this exact problem, or you want to verify an assumption before action ("am I sure about this?", "have we decided this before?"). Returns relevant memories, knowledge chunks, recent bus events, and linked priors ranked by relevance — not a wall of text. Aliases: grounding, sanity-check, verify-assumption.
In this repository142 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .claude/CLAUDE.mdCLAUDE.md
- plugin-json-schemaSkill
- readme-style-guideSkill
- releaseSkill
- scaffoldSkill
- wicked-garden-agentic-agentic-patternsSkill
- wicked-garden-agentic-architectSkill
- wicked-garden-agentic-context-engineeringSkill
- wicked-garden-agentic-frameworksSkill
- wicked-garden-agentic-performance-analystSkill
- wicked-garden-agentic-review-methodologySkill
- wicked-garden-agentic-safety-reviewerSkill
- wicked-garden-agenticSkill
- wicked-garden-agentic-trust-and-safetySkill
- wicked-garden-appsSkill
- wicked-garden-archetypeSkill
- wicked-garden-build-guideSkill
- wicked-garden-classifySkill
- wicked-garden-coreSkill
- wicked-garden-data-engineerSkill
- wicked-garden-dataSkill
- wicked-garden-deliberateSkill
- wicked-garden-demoSkill
- wicked-garden-domain-coverageSkill
- wicked-garden-domain-extractorSkill
- wicked-garden-domain-modelerSkill
- wicked-garden-domainSkill
- wicked-garden-draftSkill
- wicked-garden-editor-scaffoldSkill
- wicked-garden-engineering-api-documentarianSkill
- wicked-garden-engineering-architectureSkill
- wicked-garden-engineering-backendSkill
- wicked-garden-engineering-debuggingSkill
- wicked-garden-engineering-docsSkill
- wicked-garden-engineering-frontendSkill
- wicked-garden-engineering-integrationSkill
- wicked-garden-engineering-large-scale-migrationSkill
- wicked-garden-engineering-migration-engineerSkill
- wicked-garden-engineering-patchSkill
- wicked-garden-engineeringSkill
- wicked-garden-engineering-solution-architectSkill
- wicked-garden-engineering-system-designSkill
- wicked-garden-engineering-unit-test-qualitySkill
- wicked-garden-governed-workerSkill
- wicked-garden-jam-brainstorm-facilitatorSkill
- wicked-garden-jam-councilSkill
- wicked-garden-jamSkill
- wicked-garden-learnSkill
- wicked-garden-mcp-scaffoldSkill
- wicked-garden-mcpSkill
- wicked-garden-mem-captureSkill
- wicked-garden-mem-ingestSkill
- wicked-garden-memSkill
- wicked-garden-personaSkill
- wicked-garden-platform-auditSkill
- wicked-garden-platform-compliance-officerSkill
- wicked-garden-platform-complianceSkill
- wicked-garden-platform-errorsSkill
- wicked-garden-platform-gate-benchmark-rebaselineSkill
- The first 60 shown.
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 5 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
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 65
When invoked with a question:
- Distinctive: line 71
Step 2 — Execution: Run these, in parallel where your harness allows:
- Distinctive: line 82
Keep the top 5–10 unique signals.
- Distinctive: line 85
Step 4 — Format output: Use this shape:
- Distinctive: line 116
Never block progress. Ground is a focusing tool — absence of prior context is itself a useful signal.
- Distinctive: line 121
If you reach a decision that others should know about:
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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
wicked-garden:ground — Steer Yourself
You are uncertain. Pull what's known into focus.
When to use
- Getting mixed signals from the codebase — two things contradict each other
- About to commit to a non-obvious decision — want to know if it's been tried before
- Prior decisions might exist for this exact problem — avoid re-deriving what was deliberated
- Want to verify an assumption before taking action — gut feel needs grounding
- Picking back up on a topic after context has shifted — "let me get my bearings"
- Asking "wait, did we already decide this?" — stored memory is the answer
When NOT to use
- Routine "what does this code do" questions — use Read or Grep, they're faster
- Broad codebase exploration without a specific question — use your harness's exploration agent (or search inline)
- Fetching a specific symbol you already know exists — use the estate MCP
SearchEntitytool directly - During flow when you already have enough context — don't interrupt to re-ground
Mechanism
- Take the
questionargument (free text from the user / Claude's internal state) - Parallel query — run both simultaneously:
wicked-garden-memrecall — memories + knowledge (decisions, patterns, gotchas; estate fuses both stores, results carry source attribution)wicked-bus:query— recent bus events matching the question (last 50, filter by relevance to question terms)
- Synthesize — rank by relevance, dedupe overlap, cap at top 5–10 signals total. Priority order: memories > knowledge chunks > bus events.
- Output — dense, structured. For each hit:
- Source type:
memory,knowledge, orbus/event - One-line relevance statement (why this signal matters to the question)
- Source attribution / scope / event-id for follow-up
- Suggested follow-up (e.g., Read the cited source file)
- Source type:
- Closing pointer: "If you need more depth, open
{most relevant cited source}with Read, or runwicked-garden-memanswer for a cited synthesis." - DO NOT dump full file content — this is a focusing tool, not a firehose. The output should be skimmable in under 30 seconds.
Implementation
When invoked with a question:
Step 1 — Decompose the question into 3–5 search terms. Extract noun phrases,
named entities, and technical terms. Example: "v8 daemon projection model" →
["daemon", "projection", "v8 architecture", "state machine"].
Step 2 — Execution. Run these, in parallel where your harness allows:
Hand-off — open the wicked-garden-mem skill and run its recall action once per term ({term1}, {term2}, …); on Claude Code this is the Skill tool, on any other seat open the named skill from your catalog and carry it out inline, then continue here.
Hand-off — for recent bus events, run the wicked-bus plugin's query skill with query="{question}", limit=50 when the bus plugin is installed on your seat; skip it otherwise.
Step 3 — Rank and dedupe. Collect all results. Score by:
- Source priority (memory > knowledge chunk > bus event)
- Recency for bus events (newer = higher)
- Overlap with question terms (more term matches = higher)
Keep the top 5–10 unique signals. Drop results where two sources say the same thing — keep the higher-priority source.
Step 4 — Format output. Use this shape:
## Grounding: {question}
### What the record knows
1. [memory] {one-line relevance} — scope `{scope}`
2. [knowledge] {one-line relevance} — `{source attribution}`
3. [knowledge] {one-line relevance} — `{source attribution}`
### Recent bus activity
4. [bus/event] {event_type} @ {timestamp} — {one-line relevance}
5. [bus/event] {event_type} @ {timestamp} — {one-line relevance}
### If you need more depth
Read `{most relevant cited source}`, or `wicked-garden-mem` answer "{question}"Step 5 — If zero results from both the knowledge layer and bus, say so explicitly:
"No prior decisions or recent events found for this question. Proceeding without
grounding — consider storing the decision you reach with wicked-garden-mem (store action)."
Graceful degradation
- Knowledge layer unavailable → skip recall steps, surface bus events only, note degradation
- Bus unavailable → skip bus step, surface knowledge-layer results only, note degradation
- Both unavailable → emit: "Ground returned no context (knowledge layer and bus both unreachable). Proceeding on codebase signals only."
Never block progress. Ground is a focusing tool — absence of prior context is itself a useful signal.
After grounding
If you reach a decision that others should know about:
- Store it:
wicked-garden-mem(store action) - Emit it:
wicked-bus:emitwith the relevant event type
The value of grounding compounds when decisions are written back.
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.

