break-trace

Root-cause a reconciliation break to its source transaction or posting — follow the audit trail from the break row back to the originating entry on each side and state what differs and why. Use after gl-recon has classified a break.

Quality

87%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

87%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

Concise, highly actionable, and well-structured, with concrete MCPs, a sentence template, and a complete output schema. The main gap is the absence of a validation/verification checkpoint in a batch diagnostic workflow that feeds downstream adjust/suppress/raise-ticket actions, which caps workflow clarity.

Suggestions

Add an explicit verification step before emitting the root-cause statement — e.g., confirm the GL and subledger rows share the same key and that the differing attribute is the only meaningful delta, re-fetching or flagging ambiguity if they don't align.

Add a checkpoint for handling a break where no matching GL or subledger row is found (the trace path assumes both sides exist), so the workflow covers the not-found branch.

For batch runs over multiple breaks, note how to de-duplicate or group root-causes that share the same underlying cause before producing per-break output.

Dimension	Reasoning	Score
Conciseness	Lean and efficient with no padding and no re-explanation of concepts Claude already knows; every section — intro, 3-step trace, cause→statement, output schema, boundary note — earns its place.	5 / 5
Actionability	Fully executable guidance: names the exact MCPs (internal-gl, subledger), enumerates the precise attributes to fetch on each side, lists the attributes to diff, gives a copy-ready root-cause sentence template, and a complete output JSON schema with enum values.	5 / 5
Workflow Clarity	The 3-step sequence (pull GL → pull subledger → diff) is clear, but a batch diagnostic that drives actions like 'suppress' and 'raise-ticket' has no validation checkpoint (e.g., confirming the two sides key to the same row, or verifying the differing attribute is the true cause before emitting the root-cause statement), which caps this dimension at 3 per the batch-operation rule.	3 / 5
Progressive Disclosure	A short single-purpose skill (well under 50 lines) with no bundle files, organized into clearly headed sections (Trace path, Cause → statement, Output); per the simple-skill guidance this is a 5.	5 / 5
	Total	18 / 20 Passed

Description

87%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A well-crafted, third-person description that concretely states what the skill does and gives an explicit, specific use-trigger. Minor gains possible by listing a few more trigger synonyms, but it is already strong across all four dimensions.

Dimension	Reasoning	Score
Specificity	Names the domain and several concrete actions — 'trace to its source transaction or posting', 'follow the audit trail from the break row back to the originating entry on each side', 'state what differs and why' — but does not enumerate the full range of root-cause categories, leaving a minor coverage gap.	4 / 5
Completeness	Explicitly answers both: the 'what' (root-cause a break to its source and state what differs and why) and the 'when' ('Use after gl-recon has classified a break') with a concrete, specific trigger phrase.	5 / 5
Trigger Term Quality	Natural niche terms a practitioner would say — 'reconciliation break', 'audit trail', 'source transaction or posting', 'gl-recon' — plus the explicit trigger 'Use after gl-recon has classified a break'; good coverage but missing some common synonyms.	4 / 5
Distinctiveness Conflict Risk	A clear niche (GL reconciliation break root-causing) tied to a named upstream skill (gl-recon), with distinct triggers and minimal overlap risk against other skills.	5 / 5
	Total	18 / 20 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: anthropics/financial-services
Path: plugins/agent-plugins/gl-reconciler/skills/break-trace/SKILL.md
Commit: 69cbc81

Reviewed: 14 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.