How do I review a backend pull request for swallowed errors and silent failures?
Read every error path in the diff, not just the happy path. Flag empty catch blocks, fallbacks that hide a failure, and success responses returned after an error, and ask what the user or operator would see when each one fails.
What to bring
- The pull request diff and its description
- How the service reports errors: logs, metrics and alerts
- The behaviour the change is supposed to guarantee
A practical path
- List every try, catch, fallback and default value the diff adds or changes.
- For each one, state what happens on failure and whether anyone would notice it.
- Write review comments with the failing scenario and a concrete fix, such as rethrowing, logging with context or failing closed.
Check the result
Each comment names a line, the failure that would go unnoticed and the change that makes it visible.
Edge result and evidence
Its catalog description covers finding silent failures, inadequate error handling and fallback behaviour in code.
Also returned, in order:
- laurigates/claude-plugins@code-hidden-failures
- b-mendoza/agent-skills@review-pull-request
- Task sent
- Review a TypeScript backend pull request for silent failures, swallowed errors and correctness bugs
- Source snapshot
- Pinned SKILL.md
- Revision
- d3ebdbd5066f16a2bb8a2b8cb8af4b57c8ae324a
- Scan evidence page
- Not in the dated Edge scan evidence catalog
This is a dated search result, not a task outcome or a security pass. Results and load eligibility can change. Edge checks the selected package before a load; review its instructions and requirements.
How this page was selected
Edge is our product. This task was chosen from external Edge searches in the 30 days to 2026-10-07: 7 distinct installs searched for this kind of job and 5 of them loaded a skill. We excluded internal keys, internal installs and test clients, grouped the searches into jobs with a language model, and kept only jobs that people also search the web for. The counts describe the job group, not this exact wording. We sent the displayed prompt to find_skill on 2026-10-07 and recorded the first result and runners-up. We did not run or score the skills. No raw customer queries or identifiers are published.
Use it for your task
Give your agent the Edge setup link, then describe your own task and constraints. Ask it to find a fitting skill, load the one it chooses and finish the work. Or paste this ready-made prompt:
Set up Edge for me: fetch https://getedge.cc/SKILL.md with curl, read it, and follow it. Then find a fitting skill for my task, check its source and security state, and use it if appropriate: Review a TypeScript backend pull request for silent failures, swallowed errors and correctness bugs
Related
More answers: all task answers. Checking a skill before use: how to choose and check an agent skill and how Edge checks skills.