context-amplification-calibration

context-amplification-calibration is an agent for Claude Code from EndogenAI/dogma. It costs 0 tokens per session (4,227 once invoked), scanned A, original, Apache-2.0.

A study of a lookup table that selects which project principle an AI agent should focus on for a particular kind of task.

In plain words
What is it for?
Use it to test task-specific guidance for research, code review, commits, merges, and session closure.
Why use it?
It addresses the problem of giving every instruction equal emphasis, which can make the most relevant rules easier for the agent to overlook.

Agent for Claude Code

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

Not installable on its own: it reads a path above its own folder, which only exists inside its repository. The line is **Related**: [`docs/research/epigenetic-tagging.md`](../methodology/epigenetic-tagging.md) §3 Pattern F1–F2; [`docs/research/values-encoding.md`](../methodology.

Install

Getting it into your agent

There is no command for this one: it runs only inside a plugin, and the catalogue could not identify which plugin ships it. The source is linked below.

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 context-amplification-calibration

README.md
[![agentmods](https://agentmods.dev/badge/agents/endogenai/dogma/context-amplification-calibration.svg)](https://agentmods.dev/agents/endogenai/dogma/context-amplification-calibration)
Your own site
<a href="https://agentmods.dev/agents/endogenai/dogma/context-amplification-calibration"><img src="https://agentmods.dev/badge/agents/endogenai/dogma/context-amplification-calibration.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 4,227 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 $0.04227
Opus 5 $0.00000 $0.02114
Sonnet 5 $0.00000 $0.00845
Haiku 4.5 $0.00000 $0.00423

Measured 2d ago against content hash 7090160c50ba, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

context-amplification-calibration 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/context-amplification-calibration.md · 225 lines

How it starts

The opening of the file, as written. The whole thing — 225 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Context-Sensitive Amplification Calibration: Task-Type Axiom Activation Effectiveness

Research question: Does explicit task-type-based axiom amplification (the AGENTS.md lookup table, Phase 1 implementation) produce measurable quality improvements in agent session outcomes? Can amplification weight ratios be calibrated per task type? Date: 2026-03-10 Closes: #178 Related: docs/research/epigenetic-tagging.md §3 Pattern F1–F2; docs/research/values-encoding.md §5 OQ-VE-2; AGENTS.md §Context-Sensitive Amplification


1. Executive Summary

The context-sensitive amplification mechanism — an AGENTS.md lookup table mapping task-type keywords to the axiom that should be foregrounded at session start — was designed to address the regulatory-region gap identified in values-encoding.md §H5 and implemented in Phase 1 (2026-03-09). This synthesis provides empirical calibration evidence from analysis of ≥2 session records per task type.

Corpus: 3 task types analyzed, 6 session records examined (research, commit/merge/review, and closure/tracking task types). Evidence source: docs/sessions/, docs/plans/.

Key findings:

Task type Governing axiom Sessions analyzed Quality improvement signal Confidence
research / synthesize Endogenous-First 2 sessions ≥10%: all 4 quality gates met when amplified Medium
commit / push / review / merge Documentation-First 3 sessions ≥20%: review gate invocation rate 100% vs ~60% pre-amplification Medium
close / track / script Ambiguous amplification 1 session Signal unclear: Endogenous-First named but task was commit-type Low

Amplification weight ratios (calibrated from session evidence):

Task type keyword Primary amplify Weight ratio Secondary Ratio
research / survey / scout / synthesize Endogenous-First 1.0 ABT (source efficiency) 0.4
commit / push / review / merge / PR Documentation-First 1.0 Endogenous-First (read existing) 0.3
script / automate / encode / CI Programmatic-First (ABT) 1.0 Testing-First 0.5
agent / skill / authoring / fleet Endogenous-First 0.8 Minimal Posture 0.8
local / inference / model / cost Local Compute-First 1.0 ABT (efficiency) 0.4

Read the full file on GitHub · 225 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 · 225 lines · 0 tokens per session scan A 7090160c50ba

Subscribe to this mod's changes

context-amplification-calibration 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 4,227 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.