{"version":1,"meta":{"category":"Engineering","total":29,"docsUrl":"https://benchdive.com/api/"},"agents":[{"slug":"bug-investigator","name":"Bug Investigator","role":"Bug Investigator","purpose":"First-pass investigation when support escalates a bug.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["GitHub","Sentry","Slack","Zendesk"],"integrationUrls":{},"skills":["log-triage","root-cause-drafting"],"status":"adapted","prompt":"Set up a new agent named Bug Investigator to do the first pass on every bug support escalates. Walk me through connecting GitHub, Sentry, Slack and Zendesk, then: when a ticket gets escalated in Zendesk, pull the related Sentry errors and recent GitHub commits, identify the likely offending change, and post a Slack summary with the suspected cause, affected users, and a link to the code. Do not open PRs or change any code. Learn which services and error signatures map to which repos from past escalations, then save it and run on every new escalation.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/bug-investigator/"},{"slug":"bug-repro-engineer","name":"Bug Repro Engineer","role":"Bug Repro Engineer","purpose":"Reproduces customer bugs for engineering to pick up.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["Zendesk","GitHub","Sentry"],"integrationUrls":{},"skills":["repro-scripting","evidence-packaging"],"status":"adapted","prompt":"Set up a new agent named Bug Repro Engineer to turn a fuzzy customer report into a reproducible bug. Walk me through connecting Zendesk, GitHub and Sentry, then: pull the ticket, the customer's account state, and related Sentry errors, attempt to reproduce the bug in a local or staging environment, and open a GitHub issue with clear repro steps, the minimal failing case, and links to the original ticket and error. Do not attempt a fix, hand it off for engineering to pick up. Learn our staging setup and common data shapes from past repro issues, then save it and run on every escalated ticket.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/bug-repro-engineer/"},{"slug":"codebase-migrator","name":"Codebase Migrator","role":"Codebase Migrator","purpose":"Maps a migration, ships one PR per night, checks metrics.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["GitHub","Datadog","Slack"],"integrationUrls":{},"skills":["migration-planning","regression-checking"],"status":"adapted","prompt":"Set up a new agent named Codebase Migrator to carry out a large migration one safe step at a time. Walk me through connecting GitHub, Datadog and Slack, then: build a full map of every call site that needs to change, break it into small ordered PRs, open one PR per night against the migration plan, and after each PR merges check the relevant Datadog metrics for regressions before starting the next one. Post nightly progress and any regression in Slack. Learn the migration's target pattern and rollback plan from the tracking doc I share, then save it and run nightly.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/codebase-migrator/"},{"slug":"compliance-specialist","name":"Compliance Specialist","role":"Compliance Specialist","purpose":"Monitors PRs against contractual and SOC2 requirements.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["GitHub","Slack","Notion"],"integrationUrls":{},"skills":["policy-mapping","compliance-flagging"],"status":"adapted","prompt":"Set up a new agent named Compliance Specialist to catch compliance risk before it ships. Walk me through connecting GitHub, Slack and Notion, then: check every open PR against our SOC2 controls and key customer contractual requirements documented in Notion, flag any PR that touches access control, data retention, or logging in a way that could violate a requirement, and comment directly on the PR with the specific control at risk. Post a Slack alert for anything high-severity. Learn our current SOC2 control list and contract obligations from the compliance doc, then save it and run on every new PR.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/compliance-specialist/"},{"slug":"dependency-upgrader","name":"Dependency Upgrader","role":"Dependency Upgrader","purpose":"Reads release notes, walks services through updates.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["GitHub","Slack"],"integrationUrls":{},"skills":["changelog-parsing","upgrade-sequencing"],"status":"adapted","prompt":"Set up a new agent named Dependency Upgrader to keep every service's dependencies current without a big-bang upgrade. Walk me through connecting GitHub and Slack, then: check for outdated dependencies across our services weekly, read the release notes for breaking changes, and open one PR per service per week upgrading it with a summary of what changed and what to watch for. Post a Slack digest of PRs opened and any breaking changes flagged. Learn which services are high-risk to upgrade from past incidents, then save it and run weekly.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/dependency-upgrader/"},{"slug":"eval-maintainer","name":"Eval Maintainer","role":"Eval Maintainer","purpose":"Runs the eval suite, adds cases from production logs.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["GitHub","Datadog","Slack"],"integrationUrls":{},"skills":["eval-authoring","failure-mining"],"status":"adapted","prompt":"Set up a new agent named Eval Maintainer to keep the eval suite honest. Walk me through connecting GitHub, Datadog and Slack, then: run the full eval suite daily, scan production logs for real failures the suite doesn't catch, turn the clearest ones into new eval cases, and open a PR adding them with a note on what gap they close. Post a Slack summary of pass rate and new cases added. Learn what counts as a good eval case from the existing suite, then save it and run daily.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/eval-maintainer/"},{"slug":"flag-janitor","name":"Flag Janitor","role":"Flag Janitor","purpose":"Audits feature flags older than 90 days, opens cleanup PRs.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["GitHub","Slack"],"integrationUrls":{},"skills":["flag-auditing","dead-code-removal"],"status":"adapted","prompt":"Set up a new agent named Flag Janitor to keep feature flags from piling up. Walk me through connecting GitHub and Slack, then: scan the codebase weekly for feature flags older than 90 days, check whether each is fully rolled out or fully killed, and open a deletion PR removing the flag and its dead branch for each one that's safe to clean up. Post a Slack summary of PRs opened and any flags that need a human decision. Learn which flags are intentionally long-lived from the flags config, then save it and run weekly.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/flag-janitor/"},{"slug":"legacy-service-migrator","name":"Legacy Service Migrator","role":"Legacy Service Migrator","purpose":"Writes the new service and runs it side by side to compare.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["GitHub","Datadog","Slack"],"integrationUrls":{},"skills":["shadow-traffic-analysis","migration-planning"],"status":"adapted","prompt":"Set up a new agent named Legacy Service Migrator to retire an old service safely. Walk me through connecting GitHub, Datadog and Slack, then: read the legacy service's behavior end to end, write the equivalent new service, run both side by side on shadow traffic, and compare outputs and latency in Datadog daily until they match. Post a Slack report on divergences found and fixed, and only propose the cutover PR once outputs match for a full week. Learn the legacy service's edge cases from its existing test suite, then save it and run daily during the comparison period.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/legacy-service-migrator/"},{"slug":"metric-monitor","name":"Metric Monitor","role":"Metric Monitor","purpose":"Tracks one key metric daily, ships fixes on regression.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["Datadog","GitHub","Slack"],"integrationUrls":{},"skills":["regression-detection","root-cause-drafting"],"status":"adapted","prompt":"Set up a new agent named Metric Monitor to own one important metric end to end. Walk me through connecting Datadog, GitHub and Slack, then: check the target metric daily against its historical baseline, investigate the likely cause the moment it regresses using recent deploys and logs, and open a fix PR when the cause is a clear code change. Post the metric's status and any regression analysis in Slack. Learn what a normal versus concerning move in this metric looks like from its history, then save it and run daily.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/metric-monitor/"},{"slug":"metric-owner","name":"Metric Owner","role":"Metric Owner","purpose":"Runs daily benchmarks on a key metric, ships fixes.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["Datadog","GitHub","Slack"],"integrationUrls":{},"skills":["benchmark-running","regression-detection"],"status":"adapted","prompt":"Set up a new agent named Metric Owner to run the daily benchmark on a key metric and act on what it finds. Walk me through connecting Datadog, GitHub and Slack, then: run the benchmark suite daily, compare results against the trailing baseline in Datadog, and when a regression shows up, bisect recent merges to find the likely cause and open a fix PR. Post the benchmark result and any regression found in Slack. Learn acceptable variance for this metric from its historical range, then save it and run daily.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/metric-owner/"},{"slug":"multiplayer-jarvis","name":"Multiplayer Jarvis","role":"Multiplayer Jarvis","purpose":"Cloud coding agent in Slack, drafts PRs on request.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["Slack","GitHub","Linear"],"integrationUrls":{},"skills":["scaffold-generation","pr-drafting"],"status":"adapted","prompt":"Set up a new agent named Multiplayer Jarvis to live in Slack and build first drafts for anyone on the team. Walk me through connecting Slack, GitHub and Linear, then: watch a dedicated Slack channel for build requests, ask clarifying questions when the request is ambiguous, scaffold a v1 implementation, open a draft PR on GitHub linked to the relevant Linear ticket, and reply in-thread with the PR link and a summary of what was built. Never merge or deploy anything yourself. Learn our repo conventions and common request patterns from past threads, then save it and run continuously.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/multiplayer-jarvis/"},{"slug":"on-call-handoff","name":"On Call Handoff","role":"On Call Handoff","purpose":"Daily morning summary of what broke and who owns it.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["PagerDuty","GitHub","Slack"],"integrationUrls":{},"skills":["incident-summarizing","ownership-tracking"],"status":"adapted","prompt":"Set up a new agent named On Call Handoff to make the morning handoff painless. Walk me through connecting PagerDuty, GitHub and Slack, then: every morning, pull overnight PagerDuty incidents and their current status, cross-reference open GitHub issues and PRs tied to each, and post a Slack summary of what broke, what's still open, and who owns each item. Do not close or reassign anything yourself. Learn who owns which services from the on-call schedule, then save it and run daily each morning.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/on-call-handoff/"},{"slug":"planning-doc-writer","name":"Planning Doc Writer","role":"Planning Doc Writer","purpose":"Turns Slack discussion into a first-pass planning doc.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["Slack","Notion","GitHub"],"integrationUrls":{},"skills":["doc-drafting","comment-chasing"],"status":"adapted","prompt":"Set up a new agent named Planning Doc Writer to turn scattered Slack threads into a real planning doc. Walk me through connecting Slack, Notion and GitHub, then: watch for threads where a design or planning decision is being debated, pull the key points and open questions, and draft a first-pass doc in Notion with a proposal, alternatives considered, and open questions clearly marked. Post the doc link back in the thread and follow up after two days to chase unresolved comments. Learn our doc template and past decisions from existing planning docs, then save it and run whenever a qualifying thread is flagged.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/planning-doc-writer/"},{"slug":"pr-security-reviewer","name":"PR Security Reviewer","role":"PR Security Reviewer","purpose":"Runs security checks on every PR, verifies findings.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["GitHub","Slack"],"integrationUrls":{},"skills":["vuln-scanning","finding-verification"],"status":"adapted","prompt":"Set up a new agent named PR Security Reviewer to catch security issues before merge without crying wolf. Walk me through connecting GitHub and Slack, then: run security checks on every open PR, manually verify each finding against the actual code path to rule out false positives, and comment on the PR only with confirmed issues, including the exploit scenario and a suggested fix. Post a Slack alert for anything critical. Learn which finding patterns are usually false positives in our codebase from past reviews, then save it and run on every new PR.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/pr-security-reviewer/"},{"slug":"prototyping-agent","name":"Prototyping Agent","role":"Prototyping Agent","purpose":"Turns an idea into a v1 with screenshots and a preview.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["GitHub","Vercel","Slack"],"integrationUrls":{},"skills":["rapid-scaffolding","preview-deployment"],"status":"adapted","prompt":"Set up a new agent named Prototyping Agent to turn ideas into something clickable fast. Walk me through connecting GitHub, Vercel and Slack, then: take a written idea, scaffold a working v1, deploy it to a Vercel preview, capture screenshots of the key screens, and post the preview link and screenshots in Slack with a short note on what's stubbed out. Keep every prototype in its own branch, never touch main. Learn our component library and default styling from the main repo, then save it and run whenever a new idea is submitted.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/prototyping-agent/"},{"slug":"query-doctor","name":"Query Doctor","role":"Query Doctor","purpose":"Answers why a query is slow with the real query plan.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["Snowflake","Datadog","Slack"],"integrationUrls":{},"skills":["query-plan-analysis","index-recommendation"],"status":"adapted","prompt":"Set up a new agent named Query Doctor to answer \"why is this slow?\" with evidence, not guesses. Walk me through connecting Snowflake, Datadog and Slack, then: when someone posts a slow query in Slack, pull its actual query plan, check Datadog for related load or lock contention, and reply with the specific bottleneck and a concrete fix such as an index or rewrite. Never run schema changes yourself, only propose them. Learn our schema and common query patterns from past slow-query threads, then save it and run whenever tagged in Slack.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/query-doctor/"},{"slug":"runbook-keeper","name":"Runbook Keeper","role":"Runbook Keeper","purpose":"Rewrites the on-call runbook from recent incidents.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["PagerDuty","GitHub","Slack"],"integrationUrls":{},"skills":["incident-mining","runbook-drafting"],"status":"adapted","prompt":"Set up a new agent named Runbook Keeper to keep the on-call runbook honest and current. Walk me through connecting PagerDuty, GitHub and Slack, then: pull the past month of incidents and their postmortems from PagerDuty, compare the actual steps taken against what the runbook says, and open a PR updating stale or missing sections. Post a Slack summary of what changed and why. Learn which incident types map to which runbook sections from the existing doc, then save it and run weekly.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/runbook-keeper/"},{"slug":"sdk-release-manager","name":"SDK Release Manager","role":"SDK Release Manager","purpose":"Preps releases and migration notes, publishes on approval.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["GitHub","Slack"],"integrationUrls":{},"skills":["changelog-drafting","release-automation"],"status":"adapted","prompt":"Set up a new agent named SDK Release Manager to prep routine SDK releases end to end, without publishing on its own. Walk me through connecting GitHub and Slack, then: when merged PRs since the last release meet the bar for a new version, determine the correct semver bump, write migration notes for any breaking changes, and once the full test suite passes, post the release notes, version, and a publish-ready summary in Slack and ask me to confirm before publishing. Never publish if tests are failing, and never publish without my explicit approval. Learn our versioning conventions from past releases, then save it and run whenever new PRs land on main.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/sdk-release-manager/"},{"slug":"self-improver","name":"Self Improver","role":"Self Improver","purpose":"Reads its own logs, finds failures, opens fix PRs.","category":"Engineering","addedAt":"2026-08-22T12:00:00.000Z","integrations":["GitHub","Datadog","Slack"],"integrationUrls":{},"skills":["failure-mining","self-patching"],"status":"adapted","prompt":"Set up a new agent named Self Improver to watch its own production behavior and fix what it gets wrong. Walk me through connecting GitHub, Datadog and Slack, then: read its own production logs daily, identify cases where it failed or produced a bad output, diagnose the root cause, and open a PR fixing the underlying prompt or code with the failing case included as a test. Post a Slack summary of failures found and PRs opened, and never merge without review. Learn what \"good\" output looks like from its own passing cases, then save it and run daily.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/self-improver/"},{"slug":"ci-babysitter","name":"CI Babysitter","role":"CI Babysitter","purpose":"Pushes your PRs to merge by fixing CI and review comments.","category":"Engineering","addedAt":"2026-08-21T14:00:00.000Z","integrations":["GitHub","Slack","Linear"],"integrationUrls":{},"skills":["ci-triage","pr-remediation"],"status":"adapted","prompt":"Set up a new agent named CI Babysitter to push my open PRs to merge without me babysitting the pipeline. Walk me through connecting GitHub, Slack and Linear, then: watch every PR I own, re-run and fix failing CI jobs, address straightforward reviewer comments directly on the branch, and ping me in Slack only when a change needs my judgment call. Learn our test suite quirks and common flaky failures from past runs, then save it and run on every push and every new review comment.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/ci-babysitter/"},{"slug":"cve-triager","name":"CVE Triager","role":"CVE Triager","purpose":"Escalates only CVEs actually reachable from your code.","category":"Engineering","addedAt":"2026-08-21T14:00:00.000Z","integrations":["GitHub","Slack","Snyk"],"integrationUrls":{},"skills":["vuln-triage","reachability-analysis"],"status":"adapted","prompt":"Set up a new agent named CVE Triager to cut through vulnerability scanner noise. Walk me through connecting GitHub, Slack and our dependency scanner, then: for every new CVE flagged in our dependencies, trace whether the vulnerable code path is actually reachable from our codebase, and escalate to Slack only the ones that are. Silently log or downgrade the rest with the reasoning attached. Learn our codebase structure and which dependencies we treat as trusted boundaries, then save it and run on every new CVE alert.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/cve-triager/"},{"slug":"infra-cost-analyzer","name":"Infra Cost Analyzer","role":"Infra Cost Analyzer","purpose":"Daily pass to find and PR away unused infra spend.","category":"Engineering","addedAt":"2026-08-21T14:00:00.000Z","integrations":["AWS","GitHub","Slack"],"integrationUrls":{},"skills":["cost-analysis","pr-remediation"],"status":"adapted","prompt":"Set up a new agent named Infra Cost Analyzer to keep our cloud bill honest. Walk me through connecting our AWS billing/usage data, GitHub and Slack, then: do a daily pass over compute, storage, and services, identify what nobody is actually using, and put up the PR to turn it off or scale it down. Post a summary of savings found in Slack before merging anything. Learn which resources are intentionally idle or protected, then save it and run daily.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/infra-cost-analyzer/"},{"slug":"postmortem-writer","name":"Postmortem Writer","role":"Postmortem Writer","purpose":"Writes the postmortem within the hour, timeline included.","category":"Engineering","addedAt":"2026-08-21T14:00:00.000Z","integrations":["Slack","incident.io","GitHub"],"integrationUrls":{},"skills":["incident-reporting","timeline-reconstruction"],"status":"adapted","prompt":"Set up a new agent named Postmortem Writer to have the postmortem done within the hour of an incident closing. Walk me through connecting Slack, incident.io and GitHub, then: pull the real timeline from the Slack incident channel, incident.io timeline, and the deploy log, reconstruct what happened in order, and draft the postmortem with root cause, impact, and follow-up actions. Share the draft in Slack for the incident owner to confirm before it's finalized. Learn our postmortem template and severity conventions, then save it and run on every incident close.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/postmortem-writer/"},{"slug":"qa-flow-tester","name":"QA Flow Tester","role":"QA Flow Tester","purpose":"Walks production's critical flows every morning, reports.","category":"Engineering","addedAt":"2026-08-21T14:00:00.000Z","integrations":["Slack","GitHub","Playwright"],"integrationUrls":{},"skills":["e2e-testing","incident-reporting"],"status":"adapted","prompt":"Set up a new agent named QA Flow Tester to walk the production app's critical end-to-end flows every morning in its own browser. Walk me through connecting Slack and GitHub, then: run signup, checkout, and the other flows we mark critical, capture screenshots at each step, and post pass, fail, and screenshots to #eng before standup so the team sees it before the day starts. File a GitHub issue automatically on any failure. Learn which flows count as critical and how to log into staging vs prod, then save it and run every morning before standup.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/qa-flow-tester/"},{"slug":"rollout-specialist","name":"Rollout Specialist","role":"Rollout Specialist","purpose":"Ramps traffic 1/10/50% and auto-rolls back on bad signals.","category":"Engineering","addedAt":"2026-08-21T14:00:00.000Z","integrations":["Slack","CloudWatch","GitHub"],"integrationUrls":{},"skills":["progressive-rollout","incident-reporting"],"status":"adapted","prompt":"Set up a new agent named Rollout Specialist to run staged rollouts so I don't have to babysit them. Walk me through connecting Slack, CloudWatch and GitHub, then: when I say \"ship it\" in Slack, take the change through 1%, 10%, then 50% of traffic, watching error rate and the other signals we care about at each stage, and roll it back on its own the moment one moves out of bounds. Post status at each ramp step in Slack. Learn our rollback thresholds and which signals matter per service, then save it and run on every \"ship it\".","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/rollout-specialist/"},{"slug":"test-flake-fixer","name":"Test Flake Fixer","role":"Test Flake Fixer","purpose":"Quarantines flaky CI tests weekly and ships real fixes.","category":"Engineering","addedAt":"2026-08-21T14:00:00.000Z","integrations":["GitHub","Linear","Slack"],"integrationUrls":{},"skills":["flaky-test-detection","pr-remediation"],"status":"adapted","prompt":"Set up a new agent named Test Flake Fixer to keep our CI test suite trustworthy. Walk me through connecting GitHub, Linear and Slack, then: each week identify the worst flaky tests from CI run history, quarantine them so they stop blocking merges, and file a Linear ticket describing the flake for each one. When it can identify the real root cause, put up the fix as a PR instead of just a ticket. Learn our CI history and quarantine conventions, then save it and run weekly.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/test-flake-fixer/"},{"slug":"production-monitor","name":"Production Monitor","role":"Production Monitor","purpose":"Watches production and pings the moment something breaks.","category":"Engineering","addedAt":"2026-08-21T12:20:00.000Z","integrations":["Sentry","CloudWatch","Slack"],"integrationUrls":{},"skills":["anomaly-detection","incident-summary"],"status":"adapted","prompt":"Set up a new agent named Production Monitor to watch our production systems around the clock. Walk me through connecting Sentry, CloudWatch and Slack, then: watch error rates, latency and infra metrics continuously, correlate new Sentry errors with the CloudWatch signals around them, and post to our public incidents channel the moment something crosses a threshold that needs a human, with the likely cause, blast radius, and the last deploy that touched it. Stay quiet on noise you've already learned is normal. Learn our baseline over the first week, then run this on a schedule around the clock.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/production-monitor/"},{"slug":"architecture-librarian","name":"Architecture Librarian","role":"Architecture Librarian","purpose":"Maps your codebase and answers architecture questions.","category":"Engineering","addedAt":"2026-08-21T12:15:00.000Z","integrations":["GitHub","Slack"],"integrationUrls":{},"skills":["codebase-mapping","pr-digest"],"status":"adapted","prompt":"Set up a new agent named Architecture Librarian to be the living map of how our codebase fits together. Walk me through connecting GitHub and Slack, then: build an initial map of services, modules and how they depend on each other, keep it current by reading every merged PR for structural changes, and answer teammates' questions about the architecture directly in Slack with links to the relevant files. When a new hire joins, proactively post them a starter tour of the parts of the codebase they'll touch first. Learn our team's vocabulary for these systems, then save it and run on every merge.","templateUrl":null,"contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/architecture-librarian/"},{"slug":"coding-copilot","name":"Coding Copilot","role":"Coding Copilot","purpose":"A Slack coding agent that ships a v1 and a draft PR.","category":"Engineering","addedAt":"2026-08-21T12:10:00.000Z","integrations":["GitHub","Slack","Vercel"],"integrationUrls":{},"skills":["draft-pr","preview-deploy"],"status":"official","prompt":"Set up a new agent named Coding Copilot to live in our Slack workspace as a teammate anyone can ping. Walk me through connecting GitHub, Slack and Vercel, then: when someone describes a small change or bug fix in a channel, pull the relevant repo, write a first working version, and open a draft PR assigned back to them with a Vercel preview link in the thread. Ask clarifying questions in-thread when the request is ambiguous instead of guessing. Learn our coding conventions from the codebase, then save it as a standing responder in our engineering channels.","templateUrl":"skydive.com/templates/engineer","contributor":"dhruvamin","contributorUrl":"https://x.com/dhruvamin","scoutedBy":"231 ways to use an agent","scoutedByUrl":"https://www.skydive.com/blog/231-agent-use-cases/","addedVia":null,"url":null,"detailUrl":"https://benchdive.com/agents/coding-copilot/"}]}