profile-synthesizer

profile-synthesizer is an agent for Claude Code from zircote/human-voice. It costs 56 tokens per session (1,210 once invoked), scanned A, original, MIT.

An agent that builds a writing-style profile by combining a writer's self-assessment with computer-based analysis of writing samples. It produces a JSON profile describing the writer's voice characteristics.

In plain words
What is it for?
Use it after an interview and writing-sample analysis to merge scores, account for how reliably each trait can be self-assessed, identify distinctive features, and create a narrative summary.
Why use it?
It provides a structured way to compare what a writer says about their style with patterns found in their actual writing. It stops when required input files are missing and reports which files are absent.

Agent for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: model in frontmatter.

Part of the human-voice plugin — 2 skills, 12 commands, 3 agents, 1 hook shipped together

Good fit Use it after an interview and writing-sample analysis to merge scores, account for how reliably each trait can be self-assessed, identify distinctive features, and create a narrative summary.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/zircote/human-voice/profile-synthesizer
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.

Clone the repo
git clone --depth 1 https://github.com/zircote/human-voice

Made for: Claude Code.

Or install human-voice, the plugin that ships this one along with the rest of its 2 skills, 12 commands, 3 agents, 1 hook.

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 profile-synthesizer

README.md
[![agentmods](https://agentmods.dev/badge/agents/zircote/human-voice/profile-synthesizer.svg)](https://agentmods.dev/agents/zircote/human-voice/profile-synthesizer)
Your own site
<a href="https://agentmods.dev/agents/zircote/human-voice/profile-synthesizer"><img src="https://agentmods.dev/badge/agents/zircote/human-voice/profile-synthesizer.svg" alt="Measured on agentmods" height="20"></a>
Per session 56 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,210 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 $0.00056 $0.01210
Opus 5 $0.00028 $0.00605
Sonnet 5 $0.00011 $0.00242
Haiku 4.5 $0.00006 $0.00121

Measured 8d ago against content hash e2e92d8f9a43, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

profile-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 8d 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.

agents/profile-synthesizer.md · 105 lines

How it starts

The opening of the file, as written. The whole thing — 105 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Profile Synthesizer Agent

You synthesize voice profiles by merging self-report interview scores with computational writing analysis. You produce a complete Voice Profile JSON that captures a writer's distinctive voice characteristics.

Input

Read the following files from the session directory:

  • scores/self-report.json -- Self-report dimension scores from the interview scoring pipeline.
  • writing-samples/*-analysis.json -- Observed feature scores from computational analysis of each writing sample.
  • Any calibration data present in the session directory.

Validate that all required input files exist before proceeding. If any are missing, report the specific missing files and halt.

Tier-Weighted Merging

For each voice dimension that has both a self-report score and an observed (computational) score, apply weights based on the dimension's self-reportability tier:

Tier Self-Report Weight Observed Weight Rationale
Tier 1 0.7 0.3 High self-reportability -- writer knows this about themselves
Tier 2 0.5 0.5 Moderate self-reportability -- equal trust in both signals
Tier 3 0.3 0.7 Low self-reportability -- computational signal more reliable
Tier 4 0.0 1.0 Not self-reportable -- use observed score exclusively

The merged score for each dimension is: (SR_weight * self_report_score) + (OBS_weight * observed_score).

For dimensions with only one source (self-report only or observed only), use that source at full weight and note the single-source status in the profile metadata.

Calibration

Compare self-report versus observed scores for each dimension and classify self-awareness:

  • Delta <= 10: "high" -- Writer accurately perceives this dimension of their voice.
  • Delta <= 25: "moderate" -- Some discrepancy between perception and practice.
  • Delta > 25: "blind_spot" -- Significant gap; writer's self-perception diverges substantially from observed behavior.

Read the full file on GitHub · 105 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. 8d ago First seen · 105 lines · 56 tokens per session scan A e2e92d8f9a43

Subscribe to this mod's changes

profile-synthesizer is an agent published in the GitHub repository zircote/human-voice (10 stars, last pushed 1mo ago), licensed MIT. It adds 56 tokens to every session and 1,210 once invoked, about $0.0003 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.