Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

Autonomous Investigation Protocol

A canonical contract for AI research that runs without you: bounded question budget, search-plan gate, Fact/Inference/Assumption labels, do-not-invent lists, and a stable diffable output schema.

Data & AnalyticsIntermediate★ 6,960⑂ 827AI score 9/10Last updated: Sep 1, 2026

What it does

  • Question budget (default 3): caps clarifying questions; if nobody answers, the run proceeds with clearly labeled assumptions instead of stalling — which is what makes it schedulable.
  • Search-plan gate: before researching, the agent shows a 3-bullet plan (what to search, which source types, how fact and inference get separated) and continues unless you revise it.
  • Three-level evidence labels: every key claim is tagged Fact (checkable URL), Inference (cited evidence, your judgment on the leap), or Assumption (working guess to validate). Things it couldn't find go in an explicit gaps list.
  • Do-not-invent list: names the domain's fabrication temptations (competitors, pricing, market share, patent contents, customer wins) and forbids them; only real URLs with dates.
  • Confidence stacking: 1 channel = watch item, 2 = working hypothesis, 3+ = actionable intelligence, conflicting = dig. Announcements are treated as intent until funding, hiring, or contracts corroborate.
  • Stable schema + 4-option Final Step: section order never drifts, so run N and run N+1 diff cleanly and delta monitoring works.
  • Ethics guardrails: public, published, observable sources only — no pretexting, NDA solicitation, or terms-violating scraping (SCIP-style rule of thumb).

Who it's for

  • Product managers, strategists, and marketers who repeat competitive or market scans on a cadence.
  • Decision-makers who need to tell AI-stated facts from AI guesses.
  • Skill/agent builders who want a consistent research output contract to implement.

Examples

  1. "Run an autonomous investigation on Competitor X's move into workflow automation — this supports our Q3 roadmap bet." → 3-bullet plan → labeled findings with URLs, gaps list, and 4 next-step options.
  2. Quarterly scheduled run: same saved invocation, zero questions, identical schema — the team reads only the delta versus last quarter.
  3. Designing a new research skill: declare this protocol in References, define your domain's do-not-invent list, and lock a numbered "do not reorder" output schema.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/deanpeters/Product-Manager-Skills/HEAD/skills/autonomous-investigation/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/autonomous-investigation folder from the GitHub repo deanpeters/Product-Manager-Skills into my ~/.claude/skills/autonomous-investigation/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/deanpeters/Product-Manager-Skills.git && mkdir -p ~/.claude/skills && cp -r Product-Manager-Skills/skills/autonomous-investigation ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal.
  2. Clone the repository: git clone https://github.com/deanpeters/Product-Manager-Skills.git
  3. Create the skills directory if needed: mkdir -p ~/.claude/skills
  4. Copy the skill: cp -r Product-Manager-Skills/skills/autonomous-investigation ~/.claude/skills/
  5. Optionally copy companion skills it references (intelligence-collection-disciplines, market-landscape-scan, competitive-research-snapshot, battle-card-builder).
  6. Restart Claude Code and confirm autonomous-investigation appears in your skill list.
  7. Try it: "Use the autonomous-investigation protocol to research Competitor X's pricing changes; the decision is our next packaging revamp."
  8. Always click through the URLs behind any Fact label before acting on the findings.
View source on GitHub ↗License: NOASSERTION