CtrlK
BlogDocsLog inGet started
Tessl Logo

auto-build-repair

Headless, bounded repair of custom-code build failures in an already-generated Azure SDK PR. Thin wrapper over the shared azure-sdk-mcp:azsdk_customized_code_update engine in custom-code-only scope (editScope: CustomCode); the skill owns the iterate-until-green loop, capped by a per-language `maxIterations` read from repair-config.yml. WHEN: Copilot cloud agent runs on a release-planner Auto SDK PR labeled `auto-sdk-build-fix` that fails to build because of custom (non-generated) code. DO NOT USE FOR: full TypeSpec migrations, spec edits, API design review, manual fixing. INVOKES: azure-sdk-mcp:azsdk_customized_code_update.

67

Quality

81%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

The risk profile of this skill

SKILL.md
Quality
Evals
Security

Quality

Content

70%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.

A thorough, well-sequenced headless workflow with strong validation and feedback loops, but it pays for that thoroughness in repetition across sections and references bundle files that are not present in the skill directory.

Suggestions

Collapse the repeated "do not hand-edit / editScope: CustomCode / omit tspProjectPath / never move the pinned commit" guidance into one authoritative section and have the intro, Scope, Bounds, and Hard Rules point to it instead of restating it.

Ship the referenced co-located files (repair-config.yml, emit-repair-report.ps1, telemetry-schema.v1.json) alongside SKILL.md, or move their detail into clearly-linked reference files, so the progressive-disclosure pointers resolve.

Replace the template call-shape placeholders with one fully concrete worked example (real packagePath and a sample customizationRequest) to push actionability from mostly-executable to copy-paste ready.

DimensionReasoningScore

Conciseness

Mostly efficient and free of basic-concept padding, but the same rules ("do not hand-edit", "editScope: CustomCode", "omit tspProjectPath", "never move the pinned commit") are restated across the intro, Scope table, Bounds, Workflow, and a full "Hard Rules (recap)" section, which is noticeable redundancy.

3 / 5

Actionability

Provides a concrete call shape with parameter comments, a numbered workflow, and named artifacts (repair-config.yml, emit-repair-report.ps1, telemetry-schema.v1.json), but the MCP invocation is a template shape with placeholders and the renderer script's contents are not shown, leaving minor execution gaps.

4 / 5

Workflow Clarity

The numbered Workflow (0–5) has explicit validation checkpoints (inspect the structured result, commit only on success: true) and a clear feedback loop (re-invoke while the error set shrinks, stop on no progress or SpecChangeRequired), plus a dedicated Stop Conditions checklist.

5 / 5

Progressive Disclosure

Sections are well-organized and references are signaled one level deep, but the referenced co-located files (repair-config.yml, emit-repair-report.ps1, telemetry-schema.v1.json) are not present in the review bundle, so the disclosure structure is unverifiable/broken from the reader's perspective and significant reporting/telemetry detail is inlined.

3 / 5

Total

15

/

20

Passed

Description

92%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 tightly-scoped, third-person description that clearly states what it does, when to invoke it, and what is out of scope, with concrete actions and a named engine. Its only weakness is relying on a GitHub label and internal jargon for triggers rather than fuller natural-language phrasing.

DimensionReasoningScore

Specificity

Lists multiple concrete actions — "Headless, bounded repair of custom-code build failures", "Thin wrapper over the shared azure-sdk-mcp:azsdk_customized_code_update engine", "the skill owns the iterate-until-green loop, capped by a per-language maxIterations" — with comprehensive coverage of what it does.

5 / 5

Completeness

Explicitly answers both what (headless bounded custom-code repair via a shared engine owning the iterate loop) and when ("WHEN: Copilot cloud agent runs on a release-planner Auto SDK PR labeled auto-sdk-build-fix that fails to build"), plus a "DO NOT USE FOR" exclusion list.

5 / 5

Trigger Term Quality

Good keyword coverage including "auto-sdk-build-fix", "custom-code build failures", and "repair", but it leans on the PR label and technical jargon rather than the full set of natural phrases a user would say (e.g. "fix the SDK PR build" appears only in the body).

4 / 5

Distinctiveness Conflict Risk

Clear niche (Azure SDK custom-code build repair on a specific labeled PR via a named engine) with an explicit "DO NOT USE FOR: full TypeSpec migrations, spec edits, API design review, manual fixing" exclusion, minimizing overlap with sibling skills.

5 / 5

Total

19

/

20

Passed

Validation

93%

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

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

relative_links

Relative link issues: 4 missing

Warning

Total

15

/

16

Passed

Repository
Azure/azure-sdk-for-net
Reviewed

Table of Contents

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.