VS Code Telemetry Error Fix Guide
Trace unhandled telemetry errors to invalid-data producers instead of masking the crash site.
The skill requires data-flow tracing, rejects silently swallowing errors, and instructs truncation of user-controlled values, reducing covert-exfiltration risk. It does not specify user confirmation, rollback, least-privilege boundaries, sensitive-data redaction, or telemetry-flow limits, so points are deducted.
The guidance is internally coherent, distinguishes unknown from identifiable producers, and requests tests and compilation checks. However, there are no executable tests for this skill, abnormal-input coverage, or reproducible failure evidence. Static calibration limits this dimension to 10 or below.
The trigger, target problem type, and boundary against patching only the crash site are reasonably clear, with IPC, storage, and extension API examples. Non-fit ranges, an input/output contract, Chinese-language support, and mainland-China environment fit are not defined, so points are deducted.
The document has a clear name, description, staged workflow, positive and negative examples, and limitation-oriented guidance; repository evidence supplies MIT licensing and Microsoft attribution. The skill itself lacks versioning, changelog, maintenance ownership, update path, installation notes, and FAQs, so points are deducted.
It offers actionable investigation and repair guidance for telemetry-reported unhandled errors, with useful examples. However, it provides no standard output format, representative verified results, or static evidence of completed fixes. Static calibration therefore limits this dimension to 7 or below.
The rules, code examples, and repository tests/CI are auditable, but the supplied CI is not focused on this skill's key paths and there is no independent reproduction or third-party execution evidence. Only limited points are justified.
- Even truncated error values sent to telemetry may expose personal data, tokens, or workspace content; sensitive-data redaction should occur before truncation with an explicit field policy.
- The skill requests tests and build checks but does not define test selection, failure handling, or rollback steps for each error type; the operator must supply that verification plan.
- Telemetry dashboards, repositories, and development services may depend on overseas connectivity; the document gives no offline or mainland-China-network fallback.
What does this skill do, and when should you use it?
This skill guides investigations of unhandled errors reported by the Visual Studio Code error telemetry dashboard. It uses the error message, stack trace, hit count, and affected-user count to trace invalid data through the call stack. It directs engineers to read the error-construction code before choosing between fixing the producer, enriching telemetry context, and recognizing threshold-based conditions that may be intentional. It also requires preserving error visibility and running relevant tests and build checks after changes.
Reads the error message, stack trace, hit count, and affected-user count from a telemetry issue; examines stack frames from the bottom upward to identify data origins, expected types, and possible corruption points; searches the codebase for the error class or a unique message substring and reads the construction logic and parameter meanings; when the producer is unknown, guides enrichment at the consuming site and low-level validation function with truncated type and value diagnostics while allowing the error to remain thrown; when the producer is known, guides validation or sanitization before sending, storing, or passing data to an API; and requires relevant unit tests plus a build-task compilation check.
- A VS Code maintainer needs to locate malformed data crossing a process boundary from an error-telemetry stack trace.
- A developer is debugging type corruption originating in IPC, an extension API, storage, or user input.
- An investigator cannot identify the producer from the receiving-side stack and needs diagnostic context for the next telemetry cycle.
- A developer is fixing serialization of URI-like objects and must ensure objects are not sent as strings instead of the expected components.
- A maintainer is investigating listener-leak errors whose categories and thresholds must be understood from the construction code.
What are this skill's strengths and limitations?
- Clearly prioritizes finding the invalid-data producer over masking the failure with a guard at the crash site.
- Requires reading error-construction code, helping distinguish categories, thresholds, and actionable conditions.
- Provides a telemetry-enrichment strategy when the producer is not visible and preserves the thrown error.
- Covers validation, serialization, testing, and compilation checks as part of the repair workflow.
- It is an investigation and repair guide, not an automatic fixer or standalone command-line tool.
- The source does not provide a standalone installation process, a client-specific trigger command, or complete test commands.
- Its scope is primarily VS Code error telemetry and related code paths, not general-purpose application error handling.
- It requires access to the codebase and the ability to run relevant tests and build tasks; the source does not specify a complete runtime or platform matrix.
How do you install this skill?
The repository bundles 61 skills, but this profile covers only .github/skills/fix-errors/SKILL.md. The source does not document a standalone installation command for this skill; keep its skill directory in a location readable by an Agent Skills-compatible client. The repository as a whole is licensed under MIT.
How do you use this skill?
Invoke the skill from an Agent Skills client that can access the VS Code source tree, issue details, and telemetry information. For example: Investigate this telemetry issue: error message ..., stack trace ..., hit count ..., affected users .... First find the error-construction code, then trace the invalid-data producer through the call stack. Follow the guidance to add truncated diagnostic context when needed, fix the producer, and run relevant unit tests and build checks. The source does not specify a client-specific command or trigger syntax.