Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →automagik-dev/report
Investigate bugs comprehensively — cascade through trace, capture browser evidence, extract observability data, and prepare or explicitly create a GitHub issue with grounded findings.
Passed 13 of 18 test cases with the skill, and 7 of 18 without it.
and 8 more in the full report
8 fixed8 passed either way6 still failing0 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 8 test cases the skill fixed. Same model, same prompt, with and without the skill.
The target repository has no Sentry, PostHog, DataDog, LogRocket, or local log files configured. What message must appear in the observability section of the report?
| Checked | Without skill | With skill |
|---|---|---|
| notes No observability integrations detected in this project in the report | Missed | Passed |
The instructions your model reads when the skill loads.
Runtime syntax: invoke the plugin copy through the active runtime's owner-qualified skill selector; use a bare selector only when intentionally selecting a user-tier copy (a separately installed personal copy; Genie no longer seeds this tier). Cross-skill prose below uses bare names as portable semantic routes; the orchestrator resolves the selector for the active runtime.
Investigate a bug end-to-end: collect symptoms, run trace for root cause, capture browser evidence when available, pull observability data from project-configured tools, and prepare a GitHub issue with all findings attached. Investigation only — the deliverable is findings, never fixes; report must not modify source code. Creating the issue is a separate external write and requires explicit confirmation unless the user already asked for issue creation.
When invoked during the QA loop (after merge to dev):
.genie/wishes/<slug>/WISH.md.Criterion: "<text>" — FAIL.report → trace → fix → retest.Gather bug description (required), plus URL, error messages, and expected-vs-actual behavior when offered. If detail is missing, ask clarifying questions one at a time via native user-input surface — minimum viable input is a bug description.
The backbone of every report. Dispatch a trace subagent via the native delegation surface with a read-only brief: the symptoms, relevant files, and the expected deliverable (the trace report format — root cause file:line, evidence, causal chain, recommended correction, affected scope, confidence). The subagent notifies you with its findings as its final message; follow-ups go through native follow-up messaging. If root cause cannot be determined, note "Code investigation incomplete — trace could not determine root cause" and continue.
Requires the agent-browser CLI on PATH. Attempt when a URL was provided, a dev server is on a common port (3000, 3001, 4200, 5173, 5174, 8080, 8000), or package.json has a startable dev/start script.
| Command | Purpose |
|---|---|
agent-browser screenshot <url> [--full|--annotate] | Page evidence |
agent-browser record start <file> | Video, only for multi-step reproduction |
agent-browser profiler start / stop | Performance profile (perf bugs only) |
Also capture console errors/warnings and failed or slow network requests. Prefer screenshots over video. Unavailable → skip with note: "Browser evidence not available — no URL provided and no dev server detected."
Detect configured tools, pull recent related errors from each; integrations are independent — one failing never blocks the others.
test cases passed, vs 7 / 18 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.