verify-known-issues_skill

Overview

This skill verifies claims that something is a "known issue", "known bug", or "known limitation" before stating them as fact. It prevents hallucinated or unverified assertions about external libraries, tools, or platforms by requiring evidence, a secondary review, and citations when a claim is supported.

How this skill works

When a response would assert a known issue, the skill pauses and runs targeted web searches using multiple queries. It requires credible evidence (e.g., GitHub issues, official docs, release notes, or highly upvoted Stack Overflow answers), launches a verification sub-agent to review the findings, and formats the outcome as either a cited claim or a cautious observation.

When to use it

• Before saying "This is a known issue/bug/limitation" about any external library or tool
• When referencing reported platform behavior or incompatibilities
• When summarizing community-reported problems or regression reports
• When recommending workarounds tied to a claimed bug
• When updating documentation or release notes that mention external issues

Best practices

• Stop and do not assert a known issue until verification completes
• Run 3+ different web searches with focused queries (GitHub, docs, Stack Overflow)
• Require at least one credible source: GitHub issue, official docs, release notes, or high-vote SO answer
• Use a verification sub-agent to get a second opinion on whether evidence supports the claim
• If evidence is lacking, rephrase as an observation and list possible causes and next steps
• Cite URLs clearly when a claim is verified and summarize the supporting evidence

Example use cases

• Checking whether a reported exception in a library is a documented bug before advising a workaround
• Verifying claims about API deprecations or breaking changes prior to recommending migrations
• Confirming that a platform-specific limitation is acknowledged in release notes before including it in docs
• Assessing whether multiple user reports constitute a confirmed regression before escalating
• Turning an uncertain observation into a safe, evidence-backed statement with citations