AI Agent Hub
Back to skills
ContextLedger Knowledge Audit Ledger icon

ContextLedger Knowledge Audit Ledger

Knowledge Management Updated 2026.08.30

Paste the following prompt into your AI chat to install this skill:

Please install @user_15292d5a/yjkj-contextledger according to https://skillhub.cn/install/skillhub.md.

About this skill

Problem it solves

When engineers already have notes, docs, meeting summaries, or connector outputs, the hard part is often not retrieval but trust: sources are mixed, dates are missing, documents contradict each other, and previous long summaries may smooth away uncertainty. ContextLedger treats collected material as an audit object and produces an inspectable evidence ledger instead of another smoother summary.

How it works

It first narrows the target into an explicit claim or question, then binds 2 to 5 sources to the key conclusion and marks each as newest, oldest, undated, or likely stale. It separates direct evidence, corroborated evidence, inference, assumption, and unknown, keeps source conflicts visible rather than forcing fake consensus, and ends with the most reliable current judgment, what would change it, and the next reliable check.

Boundaries

It fits scenarios where sources already exist and need provenance or confidence review, such as policy docs, product docs, external research, and AI summary checks. If the user mainly needs ingestion, retrieval, or relationship discovery, use a connector-style tool first. If the material is thin, undated, or single-source, it reduces judgment strength and states the missing cross-source corroboration.

Use Cases

  • After reviewing multiple product docs, determine which version supports a feature claim and whether it is still valid.
  • Sort meeting notes and external research to label facts, inferences, and unsupported assumptions.
  • Check policy-doc recency risk and say whether newest or oldest sources change the current judgment.
  • Audit AI summaries for hidden source conflicts and list evidence gaps plus the next verification step.

Best For

  • Product engineers reviewing product docs who want conclusions tied to sources with stale-source flags.
  • Technical editors curating external research who need to separate direct evidence, inference, and single-source claims.
  • Legal or policy staff reviewing compliance docs who need conflicts surfaced and their impact judged.
  • Solution owners auditing AI summaries who want uncertainty preserved and the next verification step stated.