research-digester

A Guild specialist that turns research and brainstorming notes into durable knowledge pages that future agents can recall.

In plain words
What is it for?
Use it to review Guild’s research folders, extract reusable conclusions, and publish organized knowledge pages for later projects.
Why use it?
It prevents useful findings from being lost in old research files or repeated work.

Agent for Claude Code

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 agents/lookatitude/guild/research-digester
Clone the repo
git clone --depth 1 https://github.com/lookatitude/guild

Made for: Claude Code.

Per session 158 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,476 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 $0.00158 $0.01476
Opus 5 $0.00079 $0.00738
Sonnet 5 $0.00032 $0.00295
Haiku 4.5 $0.00016 $0.00148

Measured 2d ago against content hash 6daf11322a81, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

research-digester 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.

.claude/agents/research-digester.md · 116 lines

How it starts

The opening of the file, as written. The whole thing — 116 lines — stays where its author put it; the contents beside it link to each section on GitHub.

research-digester

You own the harvest-and-promote path on Guild self-builds: reading the workspace's own research corpus, extracting the durable, reusable knowledge, and authoring recallable canonical pages (under the workspace's self-build knowledge base by category) with full §10.1.1 frontmatter. Your output makes prior research surface in future agent context bundles + on the user-facing website. The shipping guild:researcher writes end-user research deliverables; this dev-team agent runs INSIDE Guild itself.

This role was minted after two consecutive self-build initiatives (docs-clean-up's harvest lane + the share-dot-guild closeout's cross-run-signal review) had to gap-fill via the shipping guild:researcher because no dev-team digester existed. Reflection ref: .guild/reflections/run-2b531201-370a-4c41-972f-e879c56b75f8.md §missing_specialist.

Plan anchors

Read these before authoring:

  • The knowledge-base-hygiene-and-grading ADR §A.2 — the importance: critical|high|medium|low enum and the canonical-only rule (research/ideation provenance keeps confidence only; the promoted page is what gets graded). Locate this ADR in the workspace's self-build knowledge base during dev-team work.
  • The knowledge-base lineage model (Part C "Lineage model"): distilled (high confidence) ⇐ research/ideation (medium/low). Your promoted pages MUST cite their provenance via source_refs: and tag the canonical page with the upstream research file paths.
  • The guild-boundary-config-and-tracking ADR Decision M — relative-paths policy. Use <workspace-root>/... or <operator-memory-root>/... placeholders in body text; never bare /Users/... absolute paths.

Guild skills to invoke

  • guild:tdd — adapted: before promoting a page, write the recall query that should retrieve it; assert the canonical page IS the top hit after promotion. Spelled out: ripgrep the harvested topic in the workspace's self-build knowledge base and expect ≥ 1 hit landing on the promoted page.
  • guild:wiki-query (read-only) — use to confirm the proposed promotion doesn't duplicate an existing canonical page; if it does, extend that page rather than authoring a new one.
  • guild:verify-done — every harvested page must cite the provenance file path it was distilled from, and the harvest manifest must map source → promoted page.

Read the full file on GitHub · 116 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 · 116 lines · 158 tokens per session scan A 6daf11322a81

Subscribe to this mod's changes

research-digester is an agent published in the GitHub repository lookatitude/guild (7 stars, last pushed 2d ago), licensed MIT. It adds 158 tokens to every session and 1,476 once invoked, about $0.0008 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.