deterministic-agent-components

deterministic-agent-components is an agent for Claude Code from EndogenAI/dogma. It costs 0 tokens per session (2,100 once invoked), scanned A, original, Apache-2.0.

A study of agent systems that separate fixed decisions, such as routing and file operations, from language-model work such as writing and summarizing.

In plain words
What is it for?
Use it to design agent workflows with deterministic routing, delegation checks, scripts, and model-based synthesis where needed.
Why use it?
It reduces unnecessary model use and helps prevent inconsistent routing when a task can be handled by a script, lookup table, or state machine.

Agent for Claude Code

Written for Claude Code: a Claude Code subagent (agents/*.md). Also seen: mentions AGENTS.md.

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/endogenai/dogma/deterministic-agent-components
Clone the repo
git clone --depth 1 https://github.com/EndogenAI/dogma

Made for: Claude Code.

Wrote 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.

agentmods badge for deterministic-agent-components

README.md
[![agentmods](https://agentmods.dev/badge/agents/endogenai/dogma/deterministic-agent-components.svg)](https://agentmods.dev/agents/endogenai/dogma/deterministic-agent-components)
Your own site
<a href="https://agentmods.dev/agents/endogenai/dogma/deterministic-agent-components"><img src="https://agentmods.dev/badge/agents/endogenai/dogma/deterministic-agent-components.svg" alt="Measured on agentmods" height="20"></a>
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 2,100 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 5.1 $0.00000 $0.02100
Opus 5 $0.00000 $0.01050
Sonnet 5 $0.00000 $0.00420
Haiku 4.5 $0.00000 $0.00210

Measured 2d ago against content hash 5ebe55e73dd7, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

deterministic-agent-components 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.

docs/research/agents/deterministic-agent-components.md · 170 lines

How it starts

The opening of the file, as written. The whole thing — 170 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Deterministic Components in LLM Agent Orchestration

Executive Summary

Pre-LLM chatbot architectures — AIML/Pandorabots, Rasa Core, Dialogflow CX, BotPress, and AWS Lex — converged independently on a structural split: a deterministic routing and decision layer, and a probabilistic content generation layer. This split is directly applicable to the EndogenAI agent fleet. A systematic mapping of the executive orchestrator workflow reveals that 63% of orchestration steps (12 of 19) are fully deterministic — script invocations, table lookups, file reads, git commands, state verifications — and require no language model inference. Extracting these steps to scripts and lookup tables reduces token burn, eliminates routing drift, and aligns with the Algorithms-Before-Tokens and Programmatic-First axioms. The recommended hybrid architecture — deterministic routing via an FSM + Delegation Decision Gate, with LLM reserved for synthesis, composition, and novel decomposition — is a direct application of the Rasa NLU/Core split to the endogenic fleet.

Hypothesis Validation

H1 — A Significant Fraction of Orchestrator Steps Are Deterministic

Verdict: CONFIRMED.

Mapping the 19 discrete steps in executive-orchestrator.agent.md (§1 Orient through §6 Session Close) against a deterministic/LLM-required binary classification yields 12 deterministic steps (63%) and 6 LLM-required steps (32%), with 1 mixed. The deterministic steps include: prune_scratchpad.py --init, reading scratchpad state, consulting the Delegation Decision Gate table, dispatching a specialist agent, confirming deliverable presence, running pre-review grep sweep, executing git commands, running prune_scratchpad.py --force, and updating issue checkboxes. The LLM-required steps — writing ## Session Start, writing ## Orchestration Plan, writing ## Pre-Compact Checkpoint, writing ## Session Summary, composing progress comments — all involve synthesis and composition that justify LLM invocation. This 63% finding aligns with BotPress production bot architecture empirics: in well-structured bots, the majority of nodes are logic nodes; LLM nodes are the minority.

Read the full file on GitHub · 170 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 · 170 lines · 0 tokens per session scan A 5ebe55e73dd7

Subscribe to this mod's changes

deterministic-agent-components is an agent published in the GitHub repository EndogenAI/dogma (2 stars, last pushed 12d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 2,100 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 2026-09-03.