reproduce-bug-report

Reproduce bug report

Use this skill when the current context is a GitHub issue, support report, Linear ticket, or user prompt describing a specific bug that may be reproduced through visible application behavior. It is primarily for UI, rendering, windowing, settings, editor, terminal-display, onboarding, or other interactive bugs where a screen recording (or screenshots) makes the result more actionable.

The parent agent should not try to manually reproduce the UI bug locally unless the user explicitly asks. Launch one or more Oz cloud agents with computer use enabled so they can run the relevant app, interact with it, and capture visual evidence.

Parent workflow

Read the bug report carefully and extract:
- reported behavior
- expected behavior
- reproduction steps, if provided
- OS, app version/build/channel, shell, feature flags, account state, or other environment constraints when relevant
- attached screenshots, videos, logs, or comments that narrow the repro path
Decide whether this skill applies:
- Use it for UI-visible bugs, interaction bugs, rendering/layout bugs, onboarding bugs, and bugs where a screen recording (or screenshots) would be useful.
- Do not use it for purely backend, CI, build, dependency, or text-only code issues unless the prompt specifically asks for visual reproduction.
- If the report requires credentials, private account state, or another capability not available to the repro environment, report that constraint clearly instead of guessing.
If the reproduction path is straightforward, launch one Oz cloud agent with computer use.
If there are multiple plausible repro paths, launch several Oz cloud agents in one run_agents batch. Give each child a distinct hypothesis or environment variant, such as:
- different OS or desktop environment
- fresh first-run state vs an already-initialized local state
- stable vs dev build
- fresh settings vs existing settings
- different shells, prompts, pane layouts, or settings toggles
If steps are incomplete, use codebase knowledge to propose likely app states and assign children to investigate those states. Do not invent facts about the original reporter's environment.
Wait for all children to report before summarizing. Distinguish confirmed reproduction, partial reproduction, non-reproduction, blockers, and untested hypotheses.

Version and app setup

Prefer reproducing against the exact app version/build/channel reported by the user when a suitable runnable artifact exists.
Do not silently substitute the latest available build when a closer reporter-matched artifact can be installed.
Prefer a released or packaged runnable artifact over a source build when that better matches the reporter's environment and the repository-specific guidance allows it.
If the exact version/build cannot be found or installed, report that clearly, explain what was attempted, and use the closest justified fallback only when it is useful for continuing the investigation.
Record the requested reporter version, the installed test version, the artifact source, and any fallback decision in the manifest and final report.

Repository-specific guidance

The consuming repository may ship a companion reproduce-bug-report-local skill. When that companion is available or referenced in the prompt, read it and apply its repository-specific scope, app setup, environment, and workflow guidance as supplemental instructions. The local companion may narrow scope or specialize setup, but it should not redefine the evidence, artifact, reporting, or safety expectations in this core skill.

Use a run_agents call shaped like this:

summary: Launching Oz cloud computer-use agents to reproduce the reported UI bug and capture a screen recording.
remote.computer_use_enabled: true
agent_run_configs:
- name: "repro-primary"
  prompt: the primary repro prompt
- name: "repro-variant"
  prompt: optional variant prompt when useful
base_prompt: the shared child prompt below

Omit extra children when they would duplicate the same steps. Omit model_id unless the user requested a specific model.

Shared child prompt

Give every child agent these shared instructions, then append the child-specific repro path or hypothesis.

Child prompt patterns

Primary repro child

Use this for a report with clear steps:

Variant child

Use this when there is a specific alternate condition worth testing:

Code-path hypothesis child

Use this when repro steps are missing or ambiguous:

Success criteria

A successful use of this skill produces:

A confirmed reproduction with a screen recording (and screenshots where useful) and exact steps, or a well-scoped non-reproduction with tested assumptions.
Clear artifact paths or attachments for visual evidence.
A concise summary of which variants were tested and which were not.
Enough environment detail for an engineer to repeat the test.
No leaked secrets, credentials, private account details, or unnecessary public comments.

Summary format

When the children finish, summarize in this structure: