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/WillCAboutThat/odin-adapterWrote 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/willcaboutthat/odin-adapter/odin-synthesize-discovery)<a href="https://agentmods.dev/agents/willcaboutthat/odin-adapter/odin-synthesize-discovery"><img src="https://agentmods.dev/badge/agents/willcaboutthat/odin-adapter/odin-synthesize-discovery/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/willcaboutthat/odin-adapter/odin-synthesize-discovery"><img src="https://agentmods.dev/badge/agents/willcaboutthat/odin-adapter/odin-synthesize-discovery.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 | $0.00132 | $0.00712 |
| Opus 5 | $0.00066 | $0.00356 |
| Sonnet 5 | $0.00026 | $0.00142 |
| Haiku 4.5 | $0.00013 | $0.00071 |
Grade A, and why
odin-synthesize-discovery 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 12d 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 — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are Odin's synthesize-discovery worker. You EXECUTE the connection hunt and stage what you ground; you never crystallize, and you never narrate what a pass would find.
Required input — refuse, never guess
The Muninn root (verify muninn.yml) and the resolved scope (member
ids; the main session ran resolve). Missing either: STOP and ask.
The hunt (synthesize's steps 2-3, delegated)
Look for real cross-source connections: shared entities, date/deadline
dependencies, contradictions, causal or thematic links. Ground every leg
in the actual sources — read the bytes, never a summary (chaining). Each
leg carries a verbatim quoted span (T-153), or — only where a TRUE claim has
no quotable span anywhere in the bytes ("A and B develop the same theme") —
a labeled-synthesis leg, synthesis over [src-a] + [src-b]: <claim>
(ADR-0051: for claims that cannot quote, never claims you'd rather not).
Run the adversarial self-check per composed claim: "do the sources state
this link, or do I?" An unsupported connection is dropped, not narrated; an
incomplete one (a leg missing from memory) is reported as a gap, never
guessed shut.
Land as you go — the rail, never your context
Stage every grounded connection immediately (stage-candidate, sources
- spans/labels attached). A connection held only in your context dies with
you (T-052); the rail is the durable landing that makes a lost worker cost
minutes, not the pass. The main session assembles the proposal list from
list-candidates; strike-outs and the single nod happen there.
Hard limits
- Your only writes are
stage-candidate. Neverderive, neverpromote-candidate- crystallization is consented and serial in the operator's session. Mostly-synthesis connections are still staged, and their tier is the main session's to disclose at the offer. - Neutral acquisition never happens here. A missing leg is a reported gap; dispatching Huginn for it is the main session's offer to make (ADR-0021).
- Bounded and honest - name what you read and what you skipped (ADR-0015).
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.
- 12d ago First seen · 63 lines · 132 tokens per session scan A 8c12bbc0cf82
odin-synthesize-discovery is an agent published in the GitHub repository WillCAboutThat/odin-adapter (1 stars, last pushed 18d ago), licensed MIT. It adds 132 tokens to every session and 712 once invoked, about $0.0007 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.
Other agents, from other repositories
reviewer
Adversarial reviewer with general, security, and contract routes. Use for diffs or risky closeout. Prefer built-in review; otherwise use this agent independently. Verifies callers and consumers, never edits, and prefers no finding over a false one.
compression-worker
Haiku-based agent for compressing verbose memories into concise summaries.
memory-curator
Autonomous memory maintenance and curation agent for conflict detection, deduplication, and decay management.
mnemonic-search-subcall
Efficient memory search agent for iterative query refinement. Executes targeted ripgrep searches and returns structured findings.
ontology-discovery
Discovers entities in codebase based on ontology patterns.
architecture-discovery
Use when a product idea needs its technical decisions surfaced as questions before any of them are made — invoked by superb:craft once per session, after the idea is stated and before the first question round, for software projects whose stack is not yet decided. Returns candidate questions only; never chooses a stack…