Explorar skills

Autonomous improvement loop for any codebase. Uses git worktrees to run every experiment in isolation — the main branch is never touched until a winning change is explicitly merged. Reads .claude/autoimprove/config.md for the measurement suite, then iterates: create worktree → propose → implement in worktree → measure → merge if improved, delete if not → log → repeat.

Detect, classify, and resolve git merge conflicts through structured analysis of conflict markers, per-file strategy selection, and post-resolution verification. Use when a branch has conflicts with its merge target, when rebasing onto an updated base, or when gh-merge detects an unmergeable PR.

Execute pending tasks for a feature — TDD-driven implementation with sub-agent isolation and progress tracking. Use when starting to build, implement, or code a planned feature, resuming partially completed work, or running the next task in a code-forge plan. Supports --repos flag for parallel implementation across multiple repositories.

Reference document describing six pre-flight checks (clean git state, not detached HEAD, gh auth, issue exists and OPEN, remote reachable, duplicate-branch warning) as pure bash exit codes with no LLM calls. Reference only (policy document; consumed by `/flow:start` Phase 0).

{{ 𝚫𝚫𝚫 }} Generate a commit message. If nothing staged, stage all changes.

Manage git worktrees for autoimprove experiments. Creates isolated worktrees for each experiment so the main codebase is never touched. Handles setup, teardown, and merging winning experiments back to main.

Analyzes a task or GitHub issue and provides effort estimation broken into subtasks with T-shirt sizing, risk assessment, and dependency mapping. Use for sprint planning or before committing to deadlines.

Use when implementation is complete and you need to merge, create a PR, or clean up — verifies tests pass, presents 4 structured integration options, executes chosen workflow, and cleans up worktrees. Pairs with code-forge:worktree.

Port a documentation-driven project to a new target language — initializes project skeleton, analyzes reference implementation, and batch-generates plans for selected features. Use when converting a project to another language, rewriting in a different language, or creating a multi-language SDK from existing feature specs.

Use before claiming work is done, fixed, or passing — requires running verification commands and confirming output before any success claim. Prevents false completion claims, unverified assertions, and "should work" statements.

Validate git conventions (commit messages, branch naming, PR format, issue linkage) by detecting project-specific rules from CLAUDE.md and settings, inferring patterns from recent history. Use when creating commits, preparing PRs, or reviewing for convention compliance. This skill MUST be consulted because convention-violating history is a defect that every future contributor must question and wor

Fetches a GitHub issue by number, analyzes the requirements, finds the relevant code, implements a fix, and optionally creates a PR. Use when working through your issue backlog.