agentleFS
PaulDanielYoungPaulDanielYoung/internal-audit-skills/skills/foundations/shared-understanding/SKILL.md

shared-understanding

Skill2 starsChanged 40 days agoMITView on GitHub

Resolve material ambiguity with the user before making a judgment or producing work. Use when missing context, assumptions, preferences, or decisions could materially change the result.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 8 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If additional material the user can provide would improve the answer, name the specific kind that would help (for example policies, …

    line 12
  2. Distinctive:

    Ask the frontier in rounds.

    line 20
  3. Distinctive:

    If the frontier is reasonably small, ask all of its questions in the same round.

    line 22
  4. Distinctive:

    Do not ask a question whose answer depends on another unresolved question in the same round.

    line 24
  5. Distinctive:

    After each round, reassess the decision tree.

    line 44
  6. Distinctive:

    If material questions remain but none can be answered until unavailable evidence or a prerequisite decision is supplied, report the blocked …

    line 50
  7. Distinctive:

    If this skill was invoked by another skill, return the settled understanding and any blockers so the invoking skill can continue supported …

    line 52
  8. Distinctive:

    If this skill was invoked directly as part of work the user already requested, continue supported work after presenting the settled …

    line 54

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Shared Understanding

Reach a shared understanding before proceeding with work that depends on unresolved decisions or assumptions.

Find the facts yourself first. Read the available files, tools, policies, procedures, prior work, and other evidence before asking the user for anything you can determine on your own. The decisions are the user's; put each material decision to them.

If additional material the user can provide would improve the answer, name the specific kind that would help (for example policies, procedures, audit reports, prior workpapers, meeting notes, etc.).

Work the decision tree

Map the unresolved questions as a decision tree. A question may depend on facts or on other decisions that must be settled first.

The frontier is the set of material unresolved questions whose prerequisites are already settled: the questions that can be answered now without guessing about another unresolved fact or decision.

Ask the frontier in rounds.

If the frontier is reasonably small, ask all of its questions in the same round. If it is large, ask the highest-leverage questions first in a manageable group while preserving dependency order.

Do not ask a question whose answer depends on another unresolved question in the same round.

For each decision question, give a recommended answer and briefly explain the reasoning. Make the recommendation specific enough that the user can accept it, modify it, or choose another option. For missing facts, ask for the fact or supporting evidence; suggest how to establish it when useful.

Format each question as shown below for its type, numbering questions in order within the round:

❓ **Q1** - **{decision question title}**

{question body, including choices where useful}

➡️ **Recommended answer** - {recommended answer and brief reasoning}
---
❓ **Q2** - **{factual question title}**

{question asking for the missing fact or supporting evidence}

🔎 **Suggested source** - {where or how to establish the fact, when useful; otherwise omit this section}

After each round, reassess the decision tree. The user's answers may settle other questions, introduce new ones, change assumptions, or move the frontier.

Know when to stop

Shared understanding is complete when no material unresolved question, decision, or assumption remains that could change the work. Summarize the resulting shared understanding when this criterion is met.

If material questions remain but none can be answered until unavailable evidence or a prerequisite decision is supplied, report the blocked questions and what would unblock them. Keep dependent work pending and continue unaffected work.

If this skill was invoked by another skill, return the settled understanding and any blockers so the invoking skill can continue supported work.

If this skill was invoked directly as part of work the user already requested, continue supported work after presenting the settled understanding and any blockers.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.