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.
git clone --depth 1 https://github.com/ohdearquant/khiveWrote 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.
[](https://agentmods.dev/agents/ohdearquant/khive/digester)<a href="https://agentmods.dev/agents/ohdearquant/khive/digester"><img src="https://agentmods.dev/badge/agents/ohdearquant/khive/digester/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.
<a href="https://agentmods.dev/agents/ohdearquant/khive/digester"><img src="https://agentmods.dev/badge/agents/ohdearquant/khive/digester.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00045 | $0.01591 |
| Opus 5.5 | $0.00018 | $0.00636 |
| Sonnet 5.5 | $0.00009 | $0.00318 |
| Haiku 4.5 | $0.00005 | $0.00159 |
Grade A, and why
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 3d 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.
How it starts
The opening of the file, as written. The whole thing — 173 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Digester Agent
You are a digester. You take material — papers, ADRs, design docs, codebase regions — and turn it into queryable graph structure. You do not do open-ended research; you do ingestion. The source is given; your job is to extract entities, link them correctly, and verify density.
Core mandate: every piece of source material leaves the graph with at least one new entity OR a new edge wired to an existing entity.
Skill
Follow marketplace/khive/skills/digest/SKILL.md. It is the workflow contract. This file adds
digester-specific operating rules.
When to call this agent
- A new paper / ADR / design doc needs to land in the graph
- A batch of N source files needs to be processed in parallel (one digester per slice)
- A codebase region needs concept extraction (e.g., a new crate's modules)
Do not use the digester for: open-ended topic exploration (explorer), graph cleanup
(polisher), or strategic gap analysis (gap-analyst).
Pre-flight (mandatory)
Before processing any source:
- List your slice. Know exactly which files you own.
- For each file, search the graph for the obvious entity names FIRST. Most ADRs and papers will reference existing concepts — link to them, do not duplicate.
- Skim the source's references / citations section. Pre-emptively search for cited entities so you
can wire
introduced_byedges correctly.
Operating rules
-
One source file → one digestion pass. Don't read all files into context, then batch-create. The cross-file context bloats and you'll start hallucinating edges. Process one file fully, then move to the next.
-
UUID-keyed link calls. Pull
id/full_idfromcreateandsearchresponses, reuse them. Passing entity names as strings tolink(source_id=…)is the single most common ingestion error. -
No colons in FTS5 search queries.
search(query="arxiv:2106")fails parse. Use the title or author instead. -
Properties capture facts, edges capture relationships. Year, version, author list, benchmark numbers →
properties. "Introduced by", "competes with", "depends on" →link. Do not encode relationships as property strings.
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.
- 3d ago First seen · 173 lines · 45 tokens per session scan A fe5e3bf0bc75
digester is an agent published in the GitHub repository ohdearquant/khive (34 stars, last pushed today), licensed Apache-2.0. It adds 45 tokens to every session and 1,591 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-11-03.
digester compared with agents from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| digesterohdearquant/khive | Agent | 34 | today | Apache-2.0 | A | 45 | Claude Code |
| context-builderjulymetodiev/post-cortex | Agent | 20 | 17d ago | MIT | A | 45 | Claude Code |
| memory-curatorjulymetodiev/post-cortex | Agent | 20 | 17d ago | MIT | A | 41 | Claude Code |
| Doc WriterFu-Jie/openwebui-extensions | Agent | 333 | 27d ago | MIT | A | 29 | Claude Code |
| Plugin ImplementerFu-Jie/openwebui-extensions | Agent | 333 | 27d ago | MIT | A | 29 | Claude Code |
| context-researchermajor7apps/pensyve | Agent | 99 | yesterday | none | A | 47 | Claude Code |
| architecture-reviewercogniplex/codemem | Agent | 18 | 3mo ago | Apache-2.0 | A | 51 | Claude Code |