A local search tool for a project's Markdown knowledge base. It finds facts, decisions, definitions, configuration details, and notes about previous fixes.

Use it to search project documentation before answering questions, investigate unfamiliar terms or errors, and retrieve relevant sections or linked notes.

It prevents the agent from guessing or answering project-specific questions from memory when the information may already be documented.

kb-search — Skill for Claude Code and Codex by BlackBeltTechnology

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.

Any agent

npx skills add BlackBeltTechnology/pi-agent-dashboard --skill kb-search

Clone the repo

git clone --depth 1 https://github.com/BlackBeltTechnology/pi-agent-dashboard

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/blackbelttechnology/pi-agent-dashboard/kb-search/github.svg)](https://agentmods.dev/skills/blackbelttechnology/pi-agent-dashboard/kb-search)

Your own site

<a href="https://agentmods.dev/skills/blackbelttechnology/pi-agent-dashboard/kb-search"><img src="https://agentmods.dev/badge/skills/blackbelttechnology/pi-agent-dashboard/kb-search/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/blackbelttechnology/pi-agent-dashboard/kb-search"><img src="https://agentmods.dev/badge/skills/blackbelttechnology/pi-agent-dashboard/kb-search.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 84 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 596 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits

NVIDIA SkillSpector pass 7 Sept 2026

How audits are shown

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.

Model	Per session	Once invoked
Fable 5.1	$0.00084	$0.00596
Opus 5	$0.00042	$0.00298
Sonnet 5	$0.00017	$0.00119
Haiku 4.5	$0.00008	$0.00060

Measured 7d ago against content hash b8a97c458a64, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

kb-search 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 7d ago.

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.

packages/kb/skill/kb-search/SKILL.md · 51 lines

How it starts

The opening of the file, as written. The whole thing — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.

kb-search — retrieve before you answer

A fast, local, zero-token FTS5 knowledge base over this project's markdown (@blackbelt-technology/pi-dashboard-kb). Retrieval is pull: you call it; nothing is auto-injected. Sub-second, deterministic, costs no model tokens — so call it freely on any uncertainty.

When to Use

Hit an unknown name / term / error string / config flag / function.
Need a past decision, convention, or "how we did X".
About to answer a factual question about this project from memory.
About to ask the user something the docs may already answer.

Procedure

Extract the key entities from the problem (names, error strings, slugs, config keys, function names).
Run: kb search "<entities>" --limit 8 --json
Read only the top 1–2 hits' full content when needed: kb get <path> --section "<heading_path>"
Still unresolved? Walk the graph from a hit: kb neighbors "<heading_path>" --depth 2 and kb backlinks "<path>".
Paraphrase miss? Lexical search is weak when your words differ from the docs' words. Reformulate once using the domain's actual terms (synonyms, the real flag/class names) and re-search. Then escalate to the user only if the KB returns nothing relevant.
Synthesize from the retrieved sections. Cite the path you used.

Pitfalls

Do NOT answer project-specific questions from memory without searching first.
Do NOT read whole files — search returns ranked sections with snippets; open full content only for the top hits.
Empty result is not a stop sign — reformulate with domain terms once, then ask.
Filter when you only want rules: kb search "<q>" --doc-type agents.

Verification

kb search returns ranked {path, headingPath, score, snippet} (lower score = more relevant).
Freshness is automatic: kb search runs an incremental reindex first unless --no-reindex.
Requires @blackbelt-technology/pi-dashboard-kb installed (kb on PATH) and a configured source (.pi/dashboard/knowledge_base.json or --source <dir>).

Read the full file on GitHub · 51 lines

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.

7d ago First seen · 51 lines · 84 tokens per session scan A b8a97c458a64

Subscribe to this mod's changes

kb-search is a skill published in the GitHub repository BlackBeltTechnology/pi-agent-dashboard (278 stars, last pushed yesterday), licensed MIT. It adds 84 tokens to every session and 596 once invoked, about $0.0004 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 2026-09-03.

Productivity Research agent agent-framework agent-orchestration agent-skills agentic agentic-ai

ask-user

Ask the human structured questions via nui HITL instead of plain chat text.

plmbr/nui · 17 tokens

starfleet-timer

Set timers instead of blocking a session with sleep/watch-loops (e.g. waiting for a GitHub Actions run): the timer fires a comms directive into the fleet at the scheduled time, and the poller injects it into the owning ship's next turn — no background process, no blocking, survives a session restart.

mpbt-hq/starfleetctl · 42 tokens

kb-search