grounded-knowledge-workflow

A workflow for answering questions from documents and other local project evidence. It can also resume a named project and keep explicit project knowledge.

In plain words
What is it for?
Use it to answer questions about your files, client context, previous research, or project state, with citations when appropriate.
Why use it?
It helps avoid mixing information from unrelated workspaces or relying on unsupported memory.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/dimosgit/grounded-knowledge-engine/grounded-knowledge-workflow
Any agent
npx skills add dimosgit/grounded-knowledge-engine --skill grounded-knowledge-workflow
Clone the repo
git clone --depth 1 https://github.com/dimosgit/grounded-knowledge-engine

Made for: Claude Code, Codex.

Per session 91 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,814 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00091 $0.01814
Opus 5 $0.00046 $0.00907
Sonnet 5 $0.00018 $0.00363
Haiku 4.5 $0.00009 $0.00181

Measured yesterday against content hash 386f3e72424d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

grounded-knowledge-workflow scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/grounded-knowledge-workflow/SKILL.md · 176 lines

How it starts

The opening of the file, as written. The whole thing — 176 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Grounded Knowledge Workflow

Use the engine as shared memory, not as a replacement for judgment. Keep this skill thin: select the semantic operation and let the MCP server and CLI enforce retrieval, project scope, citations, and writes.

Use the one-call Q&A fast path

  • For an ordinary definition, recall, explanation, or comparison, call kb.answer_and_capture exactly once with responseMode: auto and responseFormat: compact, and captureStrategy: auto.
  • Do not call kb.search or kb.get_record before it. The answer tool performs its own retrieval and grounding. Automatic retention is read-only.
  • After a successful call, return the answer, citations, capture status, tokenUsage, and timings immediately.
  • No note or review proposal is created by this fast path. Continue into maintenance only when the user explicitly asks for curation or retention.

Choose the operation

  1. For a named project or “continue where I stopped,” call kb.resume_project with the explicit projectId.
  2. For an evidence-only search request, call kb.search.
  3. For one known record, call kb.get_record by path, title, slug, or filename.
  4. For a grounded answer that may retain useful context, call kb.answer_and_capture.
  5. For project creation and administration, use the deterministic gke project CLI rather than inventing an MCP file-management workflow.
  6. When the user explicitly asks to preserve a project handoff or progress boundary, use gke checkpoint; never create one merely because a project was viewed or resumed.
  7. When the user explicitly asks to preserve or inspect a durable decision, use the full-profile decision MCP tools when available. Use the local gke decisions CLI for administration or when the MCP server is running the core profile. Local Cockpit preview/apply may append a reviewed change; the public Cockpit remains read-only.

Ground before answering

  • Check local knowledge first for questions about the user's documents, projects, prior research, or previous decisions.
  • Base factual claims on returned evidence and preserve workspace-relative citations.
  • Distinguish sourced facts from inference or recommendations.
  • If evidence is insufficient, say what is missing. Do not turn a weak match into certainty.
  • Use external research only when the user asks for it or local evidence cannot answer a question that genuinely requires current information. Keep external findings distinct from existing local knowledge.

Read the full file on GitHub · 176 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 176 lines · 91 tokens per session scan A 386f3e72424d

Subscribe to this mod's changes

grounded-knowledge-workflow is a skill published in the GitHub repository dimosgit/grounded-knowledge-engine (5 stars, last pushed 5d ago), licensed MIT. It adds 91 tokens to every session and 1,814 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

seller-research

Use when researching a merchant, storefront, marketplace seller, or merchant of record for a buying decision, especially when identity, refund terms, fulfillment, counterfeit risk, domain history, or independent buyer outcomes are uncertain.

cinderline/northcinder · 46 tokens

design-mcp-server

Design the tool surface, resources, and service layer for a new MCP server. Use when starting a new server, planning a major feature expansion, or when the user describes a domain/API they want to expose via MCP. Produces a design doc at docs/design.md that drives implementation.

cyanheads/obsidian-mcp-server · 62 tokens

api-context

Canonical reference for the unified Context object passed to every tool and resource handler in @cyanheads/mcp-ts-core. Covers the full interface, its RequestContext base, all sub-APIs (ctx.log, ctx.state, ctx.requestInput, ctx.inputs, ctx.enrich, ctx.content), and when to use each.

cyanheads/obsidian-mcp-server · 79 tokens

api-canvas

DataCanvas primitive reference — a Tier 3 SQL/analytical workspace for tabular MCP servers, backed by DuckDB. Use when registering tables from upstream APIs, running ad-hoc SQL across them, and exporting results. Covers the acquire → register → query → export flow, per-table TTL, the token-sharing pattern for…

cyanheads/obsidian-mcp-server · 85 tokens

api-testing

Testing patterns for MCP tool/resource handlers using createMockContext and Vitest. Covers mock context options, handler testing, McpError assertions, format testing, Vitest config setup, and test isolation conventions.

cyanheads/obsidian-mcp-server · 46 tokens

field-test

Exercise tools, resources, and prompts against a live HTTP server via MCP JSON-RPC over curl. Starts the server, surfaces the catalog, runs real and adversarial inputs, and produces a tight report with concrete findings and numbered follow-up options. Use after adding or modifying definitions, or when the user asks to…

cyanheads/obsidian-mcp-server · 76 tokens