init-synthesize

A synthesis agent for beastmode init, a setup process that organizes project knowledge into layered context files.

In plain words
What is it for?
Use it after writer agents finish to generate L1 summaries from L2 files and rewrite CLAUDE.md with imports and remaining guidance.
Why use it?
It turns detailed topic notes into shorter phase summaries and updates the main project guidance, so agents can use the collected context more easily.

Agent

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/bugroger/beastmode/init-synthesize
Clone the repo
git clone --depth 1 https://github.com/BugRoger/beastmode
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 961 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 $0.00000 $0.00961
Opus 5 $0.00000 $0.00481
Sonnet 5 $0.00000 $0.00192
Haiku 4.5 $0.00000 $0.00096

Measured 2d ago against content hash fb5d752bf4f9, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

init-synthesize 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.

plugin/agents/init-synthesize.md · 113 lines

How it starts

The opening of the file, as written. The whole thing — 113 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Synthesize Agent

You are the synthesis agent for beastmode init. After the writer agents have populated L2 and L3 files, you:

  1. Generate real L1 summary files from L2 content
  2. Rewrite CLAUDE.md with @imports + residual

Phase 1: Generate L1 Summaries

Read all L2 files that were just written. For each L1 file, produce a real summary.

L1 File Format

# [Phase] Context

[Summary paragraph: 2-3 sentences synthesizing all L2 topics in this phase.]

## [L2 Topic Name]
- ALWAYS [rule] — [rationale]
- NEVER [rule] — [rationale]
[max 5 rules per section]

context/<phase>/<topic>.md

L1 Files to Generate

L1 File L2 Sources
context/DESIGN.md design/product.md, design/architecture.md, design/tech-stack.md, design/domain-model.md, + dynamic
context/PLAN.md plan/conventions.md, plan/structure.md, plan/error-handling.md, plan/workflow.md, + dynamic
context/IMPLEMENT.md implement/agents.md, implement/testing.md, implement/build.md, + dynamic
context/VALIDATE.md validate/quality-gates.md, validate/validation-patterns.md, + dynamic
context/RELEASE.md release/versioning.md, release/changelog.md, release/deployment.md, release/distribution.md, + dynamic

For each L1 file:

  1. Read all L2 files in that phase directory
  2. Write a summary paragraph that captures the overall picture
  3. For each L2 file, write a section with summary + top rules + plain text path reference
  4. Include any dynamic L2 topics discovered during init

Rules for L1 Content

  • Summary paragraphs must be REAL content, not placeholders
  • If an L2 file is empty or has only placeholders, note "[Not yet populated]" for that section
  • ALWAYS/NEVER rules should be the most impactful patterns
  • L2 file paths as plain text (last line of each section), not markdown links

Phase 2: Rewrite CLAUDE.md

Read Current CLAUDE.md

If CLAUDE.md exists, read its current content. Identify:

  • Lines that are beastmode @imports (keep as-is or update)
  • Lines that match content now in L2 files (remove — redistributed)
  • Lines that don't fit any L2 topic (residual — preserve)

Read the full file on GitHub · 113 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 · 113 lines · 0 tokens per session scan A fb5d752bf4f9

Subscribe to this mod's changes

init-synthesize is an agent published in the GitHub repository BugRoger/beastmode (16 stars, last pushed 4mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 961 tokens. 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.