missingbulb
Community@missingbulb
missingbulb publishes Claudinite conventions: repo-mounted packs, checks, and task contracts governing coding-agent behavior, git discipline, CI, and lesson extraction.
Agent Skills by missingbulb
Showing 45 vetted skills indexed across 3 GitHub repositories.
missingbulbexplore-link
Extract product, market, usage, and pricing insights from a URL into a product wiki.
missingbulbwriting-wiki-pages
Enforces citation, attribution, and growth-log rules when editing product-wiki pages.
missingbulbmerge-and-ci
Manages PR-to-merge workflows with CI polling and auto-merge on GitHub.
missingbulbsnapshot-approval
Guides owner approval workflow for changed UI snapshot test baselines.
missingbulbtesting-guide
Maps the repo's test suites, requirements model, and harness documentation for test changes.
missingbulbwrite-a-saga
Author saga test cases capturing multi-step UI transitions as golden frames or animated APNG goldens.
missingbulbdeterministic-expecteds
Renders byte-stable expected screenshots using pinned clocks, faked inputs, and pixel-exact comparison.
missingbulbwrite-a-requirement-leaf
Writes backtick-numbered requirement leaves with collapsed detail blocks in requirements.md.
missingbulbchrome-store-releases
Standardizes Chrome Web Store release pipelines across extension repositories using vendored GitHub Actions workflows.
missingbulbextension-host-permissions
Requests runtime host permissions for Chrome extension content scripts on arbitrary sites.
missingbulbweb-speech-io
Wire browser speech recognition and text-to-speech APIs into web applications.
missingbulbnode-test-discovery
Validates node --test invocations so test paths resolve to real files instead of silently running zero tests.
missingbulbfile-placement
Evaluates file placement using a reference-distance metric against the repository folder tree.
missingbulbauthoring-agent-docs
Guides writing CLAUDE.md and instruction files that coding agents follow reliably.
missingbulbwriting-repo-scanning-checks
Guides writing repo-scanning checks that select files, strip comments, and validate against real sources.
missingbulbcommitting
Enforces one-concern-per-commit discipline with issue references and explicit file staging in git.
missingbulbci-performance-evaluation
Diagnose where CI pipeline time goes and validate which optimizations are worth shipping.
missingbulbwriting-handover-issues
Writes GitHub issue checklists that a human executes step by step.
missingbulbsearching-for-a-tool
Guides ToolSearch queries for finding deferred harness tools by name.
missingbulbdo-later
Files deferred changes as chained GitHub issues blocked behind current work.
missingbulbbug-investigation
Diagnose bug root causes using observation-driven investigation methods.
missingbulbverify-in-production
Files deferred production verification issues for merged changes that cannot be observed immediately.
missingbulbfetching-from-the-web
Guides web fetching decisions for exact bytes, 403 responses, and egress blocks.
missingbulbrepo-text-sweeps
Guides grep and sed sweeps for renames and path relocations across repositories.
Frequently Asked Questions About missingbulb
FAQPage SchemaWhat tasks can I perform using missingbulb's skills?▼
You can bootstrap and adopt Claudinite packs, enforce commit and merge discipline, evaluate CI runtime, write migration plans and handover issues, investigate bugs, author agent instruction docs, convert pack prose into repo-scanning checks, and process research paper PDFs with vessel-overlay review.
Who are missingbulb's skills designed for?▼
They target engineers running coding agents inside repositories — maintainers who need enforceable conventions for commits, tests, CI, file placement, and agent instruction docs, plus teams extracting reusable lessons from repo activity and conversation logs into local packs.
How do the skills load and run in practice?▼
Skills load conditionally via native frontmatter triggers: file-edit path globs (e.g. **/CLAUDE.md, **/*.test.*), tool-call matchers (git commit, WebFetch, ToolSearch, GitHub merge), and prompt patterns like '/do-later' or 'LGTM', so guidance appears exactly when the matching action occurs.
What prerequisites or dependencies do these skills assume?▼
They assume a repo running the Claudinite harness with mounted packs, hooks, and declared checks, plus GitHub integration (MCP issue and merge tools), git, and GitHub Actions. The paper-intake and vessel-overlay skills assume a research repo with a references/ folder and a segmentation pipeline.
How do the growth and revalidation skills keep pack rules accurate?▼
extract-from-activity and extract-from-conversations mine commits, PRs, and session logs for lessons; prose-to-checks converts testable prose into checks; growth-dedup prunes local items the canon now covers; revalidating-rules re-probes environmental claims like tool contracts and platform behavior for staleness.