Run any Skill in Manus with one click

Get Started

run-and-analyze

Stars2

Forks1

UpdatedFebruary 25, 2026 at 00:23

Execute experiment run.sh, run analysis script, parse output, and interpret results

Installation

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

Run Skill in Manus

Source

inference-sim

inference-sim/sdlc-plugins

View GitHub Repository View Creator Repositories

Download

Run Skill in Manus

Related occupationsSOC

Based on SOC occupation classification

Software Quality Assurance Analysts and TestersComputer and Mathematical Occupations·SOC 15-1253

SKILL.md

readonly

name	_run-and-analyze
description	Execute experiment run.sh, run analysis script, parse output, and interpret results
user-invocable	false
allowed-tools	["Bash(./hypotheses/)","Bash(cd && ./run.sh)","Bash(python3 )","Bash(node )","Read","Write","Edit","Glob","Grep"]

Run and Analyze Experiment

Execute the experiment and produce interpreted results.

Invocation

Skill(_run-and-analyze "<hypothesis_dir>" "<claim>")

Arguments:

hypothesis_dir — path to hypothesis directory (e.g., hypotheses/h3-caching-latency)
claim — the original hypothesis claim (used for interpretation context)

Step 1: Run the Experiment

Bash: cd <hypothesis_dir> && ./run.sh

Use a 5-minute timeout.

Handling Failures

If run.sh fails:

Read the error output — identify the problem
Auto-fix if possible (build errors, bad flags, import errors) by editing run.sh
Re-run. Max 3 retry attempts.
If still failing, report error and stop with VERDICT: Failed.

Step 2: Verify Output Files

Check <hypothesis_dir>/output/ for config_a.txt and config_b.txt. If missing or empty, report and stop.

Step 3: Run Analysis Script

Bash: cd <hypothesis_dir> && python3 analyze.py

Auto-fix parsing issues if needed (max 3 attempts). Capture analysis output.

Step 4: Parse and Compare

Read output files, extract metrics, build comparison table:

Metric name, Config A value, Config B value, delta, percent change

Step 5: Interpret

Confirmed: Dependent variable moved in predicted direction by meaningful amount
Refuted: Did NOT move as predicted, or opposite direction
Inconclusive: Mixed results, tiny effect, questionable data

Explain in 2-3 sentences.

Output

Return this structured text to the calling skill:

VERDICT: [Confirmed|Refuted|Inconclusive|Failed]
SUMMARY: [2-3 sentence interpretation]
METRICS_TABLE: [formatted comparison table]
ANALYSIS_OUTPUT: [raw output from analyze script]

More from this repository

same repository

formulate-hypothesis

inference-sim/sdlc-plugins

Formulate a single testable hypothesis by scanning the project for untested behaviors, gaps, and claims

2026-02-252

hypothesis-test

inference-sim/sdlc-plugins

Guided hypothesis-driven experimentation. Choose a project, generate hypotheses, select which to test, run experiments in parallel or sequentially, document findings.

2026-02-252

research-ideas

inference-sim/sdlc-plugins

Generate iteratively-reviewed research ideas from a problem statement. Fully guided - no arguments required.

2026-02-252

review-plan

inference-sim/sdlc-plugins

Send Claude plan to an LLM for external technical review

2026-02-252

summarize-problem-context

inference-sim/sdlc-plugins

Create research document with problem statement and context from multiple sources (repositories, papers, files)

2026-02-252

document-findings

inference-sim/sdlc-plugins

Populate FINDINGS.md with results and analysis, update HYPOTHESIS.md status

2026-02-252

name	_run-and-analyze
description	Execute experiment run.sh, run analysis script, parse output, and interpret results
user-invocable	false
allowed-tools	["Bash(./hypotheses/)","Bash(cd && ./run.sh)","Bash(python3 )","Bash(node )","Read","Write","Edit","Glob","Grep"]

Run and Analyze Experiment

Execute the experiment and produce interpreted results.

Invocation

Skill(_run-and-analyze "<hypothesis_dir>" "<claim>")

Arguments:

hypothesis_dir — path to hypothesis directory (e.g., hypotheses/h3-caching-latency)
claim — the original hypothesis claim (used for interpretation context)

Step 1: Run the Experiment

Bash: cd <hypothesis_dir> && ./run.sh

Use a 5-minute timeout.

Handling Failures

If run.sh fails:

Read the error output — identify the problem
Auto-fix if possible (build errors, bad flags, import errors) by editing run.sh
Re-run. Max 3 retry attempts.
If still failing, report error and stop with VERDICT: Failed.

Step 2: Verify Output Files

Check <hypothesis_dir>/output/ for config_a.txt and config_b.txt. If missing or empty, report and stop.

Step 3: Run Analysis Script

Bash: cd <hypothesis_dir> && python3 analyze.py

Auto-fix parsing issues if needed (max 3 attempts). Capture analysis output.

Step 4: Parse and Compare

Read output files, extract metrics, build comparison table:

Metric name, Config A value, Config B value, delta, percent change

Step 5: Interpret

Confirmed: Dependent variable moved in predicted direction by meaningful amount
Refuted: Did NOT move as predicted, or opposite direction
Inconclusive: Mixed results, tiny effect, questionable data

Explain in 2-3 sentences.

Output

Return this structured text to the calling skill:

VERDICT: [Confirmed|Refuted|Inconclusive|Failed]
SUMMARY: [2-3 sentence interpretation]
METRICS_TABLE: [formatted comparison table]
ANALYSIS_OUTPUT: [raw output from analyze script]