glean-result-vetting

An always-on rule for checking results returned by Glean, a tool for searching an organisation’s documents and knowledge. It checks whether each result is relevant, recent, and from a reliable source.

In plain words
What is it for?
Use it whenever presenting Glean search results to decide which documents to include and how strongly to trust them.
Why use it?
It helps prevent search answers from being padded with keyword matches, outdated documents, or informal notes. It also requires useful source details such as the title, link, update date, and confidence.

Cursor rule

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 rules/gleanwork/cursor-plugins/glean-result-vetting
Clone the repo
git clone --depth 1 https://github.com/gleanwork/cursor-plugins
Per session 217 This file is loaded in full into every session.
When invoked 217 The same file — it is already loaded in full.
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 $0.00217 $0.00217
Opus 5 $0.00109 $0.00109
Sonnet 5 $0.00043 $0.00043
Haiku 4.5 $0.00022 $0.00022

Measured 2d ago against content hash 8149223909d1, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

glean-result-vetting 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.

glean/rules/glean-result-vetting.mdc · 23 lines

What it actually says

Glean Result Vetting

When presenting results from Glean MCP tools, always vet before showing:

Relevance: Does this actually answer the question, or just contain matching keywords? Reject keyword coincidences.

Freshness: Is this current?

  • ✅ <6 months: include with confidence
  • ⚠️ 6-12 months: include with staleness note
  • ❌ 12+ months: exclude unless no alternatives (with strong warning)

Authority: How reliable is the source?

  • Official (RFCs, approved specs) > Semi-official (team wikis) > Informal (Slack, personal notes)

Quality over quantity: 3 vetted results beats 10 unvetted mentions. "Nothing relevant found" is a valid and useful answer — never pad with weak results.

Always include: document title, URL, last updated date, and confidence level when presenting Glean results.

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 · 23 lines · 217 tokens per session scan A 8149223909d1

Subscribe to this mod's changes

glean-result-vetting is a cursor rule published in the GitHub repository gleanwork/cursor-plugins (4 stars, last pushed 12d ago), licensed MIT. It adds 217 tokens to every session, about $0.0011 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.