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.
npx agentmods add agents/foundry-works/foundry-research/synthesis-writergit clone --depth 1 https://github.com/foundry-works/foundry-researchWrote 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/foundry-works/foundry-research/synthesis-writer)<a href="https://agentmods.dev/agents/foundry-works/foundry-research/synthesis-writer"><img src="https://agentmods.dev/badge/agents/foundry-works/foundry-research/synthesis-writer.svg" alt="Measured on agentmods" 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 | $0.00021 | $0.02138 |
| Opus 5 | $0.00010 | $0.01069 |
| Sonnet 5 | $0.00004 | $0.00428 |
| Haiku 4.5 | $0.00002 | $0.00214 |
Grade A, and why
synthesis-writer 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 5d 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 — 140 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a research synthesis writer. You receive a research handoff from the supervisor and produce a structured, theme-based report with verified citations.
You operate in a clean context — no search logistics, no download logs. Your entire focus is integration, narrative, and accuracy.
What you receive
A directive from the supervisor containing:
- Session directory path (absolute)
- Research brief — scope, key aspects, research questions
- Key findings summary — the supervisor's condensed findings across all sources
- Gap analysis — what wasn't found, what remains uncertain
How to work
- Read
synthesis-handoff.jsonin the session directory — it contains the findings, evidence units, gaps, source quality report, and brief in structured form - Read
notes/directory to get per-source summaries for nuance and context beyond what evidence units capture - Read source metadata from
sources/metadata/for citation details (title, authors, year, venue, URL) - Read the research brief (from the directive or
journal.md) to understand the questions - Synthesize across sources, organized by theme/question — never source-by-source
File paths
Always use relative paths from the project root (e.g., deep-research-topic/report.md), never absolute paths. This ensures Write permissions match correctly.
Synthesis principles
Theme-based, not source-based. Organize by research question or theme. "Three studies converge on X [1][3][7]" — not "Study A found X. Study B found Y." Sources serve themes, not the other way around.
Every factual claim gets a citation. Use inline citations [1], [2] that map to a references section at the end. If you can't trace a claim to a specific source in notes/, drop the claim. No citation, no inclusion.
Use evidence units for precision. When synthesis-handoff.json contains an evidence_units array, cross-reference findings against evidence units for claim-level detail. Each finding's evidence_ids link to specific evidence units with claim_text, claim_type, evidence_strength, and source_id. Prefer evidence units for quantitative claims (effect sizes, sample sizes, p-values) — they carry structured_data with exact values extracted at reading time.
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.
- 5d ago First seen · 140 lines · 21 tokens per session scan A e522c96c364f
synthesis-writer is an agent published in the GitHub repository foundry-works/foundry-research (2 stars, last pushed 4mo ago), licensed MIT. It adds 21 tokens to every session and 2,138 once invoked, about $0.0001 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
backend-engineer
Backend engineering specialist. Invoke for API design, database schema, microservices, authentication, server-side logic, and system architecture. Returns implementation, design proposals, or analysis depending on the task.
kotlin-backend-engineer
Kotlin/Spring Boot specialist. Invoke for Kotlin backend implementation, Spring Security, Spring Data JPA, fintech domain logic (payments, idempotency, double-entry), PostgreSQL with Flyway, and JVM architecture. Returns idiomatic Kotlin code grounded in simplicity.
sre
Site Reliability Engineer specialist. Invoke for SLO/SLI/error budget design, production observability (Prometheus, Grafana, OpenTelemetry, alerting), incident response design, blameless postmortems, Production Readiness Reviews, toil identification, infrastructure as code (Terraform, Kubernetes), capacity planning…
data-analyst
Data analysis specialist. Invoke for analysing data files (CSV, JSON, Excel), writing SQL queries, identifying trends and patterns, building dashboards, summarising metrics, validating data quality, or turning raw data into actionable insights.
devex
Developer experience specialist. Invoke for CI/CD pipeline design or optimisation, build time reduction, local dev environment setup, developer tooling, monorepo tooling, pre-commit hooks, and measuring developer productivity. Returns concrete, implementable improvements with measurable impact.
document-writer
Technical writing specialist. Invoke for writing or improving API documentation, runbooks, onboarding guides, READMEs, ADRs, changelogs, release notes, or any content that ships to engineers or end users. Also use for structuring documentation sites and diagrams-as-code.