analyze-results

An automated runbook for cleaning and evaluating results from crowdsourced speech-quality studies. It supports tests such as ACR, where listeners give a rating, and related speech and echo tests.

In plain words
What is it for?
It helps run the result parser, check data quality, calculate average scores for each recording and worker, and create a file for running the same analysis again.
Why use it?
It turns raw answers into checked results and helps make the analysis repeatable instead of relying on manual processing.

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/microsoft/p.808/analyze-results
Clone the repo
git clone --depth 1 https://github.com/microsoft/P.808
Per session 52 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,840 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 $0.00052 $0.02840
Opus 5 $0.00026 $0.01420
Sonnet 5 $0.00010 $0.00568
Haiku 4.5 $0.00005 $0.00284

Measured 2d ago against content hash 741ed80ec31e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

analyze-results 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.

.github/agents/analyze-results.agent.md · 269 lines

How it starts

The opening of the file, as written. The whole thing — 269 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Analyze subjective test results

Use this runbook when asked to analyze, parse, or evaluate results from a completed subjective speech quality test (ACR, DCR, CCR, P.835, P.804, echo impairment, or personalized P.835).

Trigger phrases: "analyze results", "parse results", "evaluate the study", "process the answers", "run result parser".

Platform and shell adaptation

Code examples use PowerShell on Windows (\ paths). Adapt for other OS/shells: replace PowerShell cmdlets with equivalents, use python3 if needed, convert paths. Replace REPO_ROOT with the actual absolute path of this repository.

Mandatory pre-check

Before running anything:

  1. Read AGENTS.md and .github\copilot-instructions.md.
  2. Confirm this is an analysis task, not creation. For study creation, use the create-study agent instead.

Environment prerequisites

Verify once at the start:

  1. Python deps: pip install -r requirements.txt --quiet in src\.

Inputs the agent must collect

Do not guess these values if they are missing:

  1. Test method: one of acr, dcr, ccr, p835, p804, echo_impairment_test, pp835.
  2. Result parser config file (*_result_parser.cfg): generated by master_script.py during study creation. Located in the project output directory.
  3. Answers CSV (Batch_XXX.csv): exported from the crowdsourcing platform (AMT) or HIT App server. Contains worker responses.
  4. Prolific demographic CSV (optional): prolific_demographic_export_*.csv — only needed if the study was run on Prolific via HIT App server.
  5. Payment per session (Prolific only): the reward paid to a participant per session/HIT (e.g. 2.10). Prolific does not include the reward in its export, so the parser cannot compute payment-per-hour statistics without it. Ask the user for this value whenever the study was run on Prolific; pass it to the parser as --payment_per_session. Not needed for AMT (the reward is already a column in the AMT batch).

Read the full file on GitHub · 269 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 · 269 lines · 52 tokens per session scan A 741ed80ec31e

Subscribe to this mod's changes

analyze-results is an agent published in the GitHub repository microsoft/P.808 (230 stars, last pushed 14d ago), licensed MIT. It adds 52 tokens to every session and 2,840 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 2026-08-30.