charley-skills

AGENTS.md: create, audit, or maintain the file. Use when the user mentions AGENTS.md, project memory, instruction budget, or progressive disclosure of agent instructions.

Persist gotchas, preferences, or a repeated workflow from recorded agent history into AGENTS.md or a new skill.

Take over an unfinished coding session recorded by AgentsView — from Codex, Claude Code, Cursor, Copilot, Grok, or another machine — by reconstructing a short handoff and verifying it against the current repository.

Use when managing a Node.js project's dependencies or scripts with aube (https://aube.sh/), or migrating a project from pnpm, npm, or bun to aube — including its lockfile, CI, lifecycle-script jail, and Dependabot handling.

Use when backfilling unit test coverage on an existing, under-tested codebase — detecting the project's test framework across any language, generating tests for coverage gaps, and validating that they build, run under CI, and actually fail on broken code. For interactive feature-first development, use the tdd skill instead. Out of scope: integration, end-to-end, browser, and performance tests.

Use ONLY when a git repository is hosted on GitHub to create, structure, or link GitHub epics, sub-issues, and dependency relationships.

Use ONLY when a git repository is hosted on GitLab to create, structure, or link GitLab epics, sub-issues, and issue relations.

mise: set up a project's toolchain in mise.toml, migrate Makefile or npm scripts to tasks, or wire mise into CI or containers.

Use when preparing, creating, updating, or reviewing pull requests — enforcing pre-PR checks, atomic commits, dependency release links, and issue linkage (GitHub primary, non-GitHub referenced).

Set up a repository so agents can file tickets, open pull or merge requests, and verify their own changes.

Tech-lead: split implementation into slices, brief a subagent, isolate independent slices in git worktrees, and accept the result in the primary session. Use when delegating implementation to subagents or running parallel worktrees.

Use when cleaning up local git commit history before review or merge — squashing WIP/fixup noise, reordering, rewording, splitting mixed-concern commits, dropping noise, or fixing unsigned commits.

Explicitly record something worth remembering, deciding scope (global vs. project) and tier (short-term candidate vs. long-term durable), then writing it. Use on `/to-memory`, "remember this", or `$to-memory`. Autonomous knowledge capture after solving a problem is agents-md's Prevent Recurrence, not this skill.

Record a website walkthrough video with auto-zoom on clicks. Use when the user asks for a walkthrough, when a PR, MR, or issue needs a recording of a web flow, or on /to-walkthrough-video.

Use when the user wants browser UI e2e as a Playwright Test spec — converting a completed webwright run, or unblocking a missing Playwright toolchain or webwright run first. Not for API-level end-to-end tests.