query

A question-answering command for a curated wiki, a collection of linked reference pages. It searches the wiki map and pages, follows related links, and cites claims with links to the source pages.

In plain words
What is it for?
Use it to ask natural-language questions about the wiki, trace supporting pages, identify missing knowledge, and optionally save a new answer to the wiki inbox.
Why use it?
It gives answers from the approved knowledge base instead of silently relying on unreviewed source documents.

Command

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 commands/capitalone/context-specs/query
Clone the repo
git clone --depth 1 https://github.com/capitalone/context-specs
Per session 59 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,520 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 $0.00059 $0.01520
Opus 5 $0.00030 $0.00760
Sonnet 5 $0.00012 $0.00304
Haiku 4.5 $0.00006 $0.00152

Measured 2d ago against content hash 42737b15801f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

query 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 2d 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.

skills/human-loop/wiki-init/assets/commands/query.md · 113 lines

How it starts

The opening of the file, as written. The whole thing — 113 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/query

Read CLAUDE.md first for the layered structure and hard rules. Critical rule for this command: search wiki/ only — never read from raw/. Synthesis-at-ingest is the whole point of the wiki; if a question can't be answered from wiki/, the right answer is "ingest more sources," not "let me go read the raw docs."

Arguments: $ARGUMENTS

The argument is the natural-language question. If $ARGUMENTS is empty, ask the user for their question and stop.

Step 1 — Topology pass (read MOC first)

Before grepping, read wiki/MOC.md end-to-end. MOC is the curated map of the vault — it tells you which clusters/sections exist and which pages are considered canonical entry points for each area. Note which sections are most relevant to the question; you'll use them to weight grep results in step 2 and to notice gaps (e.g. "MOC has no section that covers this question — that itself is a finding").

If MOC obviously hasn't been updated for pages you encounter later (e.g. step 2 surfaces a clearly-relevant page that MOC doesn't link), flag this in your final answer as a lint-worthy finding so the user knows MOC is drifting.

Step 2 — Find the seed pages

Identify keywords/concepts in the question. Use Grep over wiki/ (NOT raw/) to find candidate pages. Match against:

  • Filenames (slugs).
  • H1 titles.
  • aliases: frontmatter values.
  • First paragraph / Definition section text.

Score candidates by:

  1. How directly the page addresses the question (keyword density, definition match).
  2. Whether MOC lists it under a section the question maps to (boost).

Take the top ~5 as seed pages.

For each seed page, read it. Follow outgoing [[wikilinks]] to add neighbors to the working set. Then follow links one more hop. Cap the working set at ~15 pages — beyond that, the question is too broad, narrow it.

While reading, note:

  • Which pages contain claims relevant to the question.
  • Which sources ([[<source-slug>]]) those claims cite.
  • Any contradictions or open questions surfaced by the pages.

Read the full file on GitHub · 113 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.

  1. 2d ago First seen · 113 lines · 59 tokens per session scan A 42737b15801f

Subscribe to this mod's changes

query is a command published in the GitHub repository capitalone/context-specs (41 stars, last pushed 8d ago), licensed Apache-2.0. It adds 59 tokens to every session and 1,520 once invoked, about $0.0003 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 2026-08-30.