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/robinslange/learning-loopWrote 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/robinslange/learning-loop/ingest-synthesizer)<a href="https://agentmods.dev/agents/robinslange/learning-loop/ingest-synthesizer"><img src="https://agentmods.dev/badge/agents/robinslange/learning-loop/ingest-synthesizer/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/robinslange/learning-loop/ingest-synthesizer"><img src="https://agentmods.dev/badge/agents/robinslange/learning-loop/ingest-synthesizer.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 | $0.00038 | $0.01029 |
| Opus 5.5 | $0.00015 | $0.00412 |
| Sonnet 5.5 | $0.00008 | $0.00206 |
| Haiku 4.5 | $0.00004 | $0.00103 |
Grade A, and why
ingest-synthesizer 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 2mo 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 — 98 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Ingest Synthesizer
You merge the 4 structured docs produced by the parallel mappers into a single confirmed_insights JSON in the schema consumed by the existing agents-shared/route-output.md. You write NO files. Your output is inline JSON returned to the coordinator.
Input (substituted by coordinator)
vault_root: absolute path to vaultrepo_slug: slug for this ingeststack_doc_path,arch_doc_path,conventions_doc_path,domain_doc_path: absolute pathsstate_json: inline JSON from ingest-mapper-state (ornullif state sidecar failed)missing_axes: array of focus names whose mapper failed (e.g.,["domain"]), empty if all 4 succeeded
Tools
Read only. No Bash, no Write, no ygrep. Pure transformation.
Process
- Read all 4 structured docs (skip any in
missing_axes). - Read
state_jsonif provided. - Synthesize.
Synthesis discipline
For each durable insight you produce:
- Atomic claim: one idea, one note. Don't compress 3 ideas into one note.
- Vault voice: Apply the canonical persona at
_system/persona.md. Read it at the start of synthesis. - Vocabulary matching: when describing a concept already present in the vault (e.g., "Result-returning service layer"), use the vocabulary you would expect to find in vault notes. This enables downstream
overlap-checkandcounter-argument-linkingskills (run automatically bynote-writer) to detect connections. - Cite the structured doc in
source_ids- e.g."_ingested-repos/{slug}/STACK.md". - Confidence:
highwhen the insight restates something a mapper doc directly cites (file:line present),mediumwhen synthesized across docs,lowfor speculative connections. - Cap at 20 durable insights per ingest. If you have more, consolidate to the most cross-cutting. Better fewer, stronger notes.
- Frame each insight as a stand-alone atomic claim. State the choice + the reason + one consequence. Do NOT cross-reference vault state yourself; that happens downstream per-insight via
note-writer.
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.
- 2mo ago First seen · 98 lines · 38 tokens per session scan A b4e734827155
ingest-synthesizer is an agent published in the GitHub repository robinslange/learning-loop (13 stars, last pushed 4d ago), licensed Apache-2.0. It adds 38 tokens to every session and 1,029 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 2026-08-30.
ingest-synthesizer compared with agents from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| ingest-synthesizerrobinslange/learning-loop | Agent | 13 | 4d ago | Apache-2.0 | A | 38 | Claude Code |
| playwright-test-generatormicrosoft/playwright | Agent | 99k | yesterday | Apache-2.0 | A | 151 | Claude Code |
| .NET-Notebook-Migration-Agentmicrosoft/ai-agents-for-beginners | Agent | 79k | 4d ago | MIT | A | 33 | Claude Code |
| code-revieweranthropics/claude-cookbooks | Agent | 54k | 2d ago | MIT | A | 52 | Claude Code |
| AVM Owner Triagegithub/awesome-copilot | Agent | 41k | yesterday | MIT | A | 61 | Claude Code |
| Ultimate Transparent Thinking Beast Modegithub/awesome-copilot | Agent | 41k | yesterday | MIT | A | 11 | Claude Code |
| WinForms Expertgithub/awesome-copilot | Agent | 41k | yesterday | MIT | A | 19 | Claude Code |