Coordinate investigation, fixes, tests, and commits for bugs, Claude Code hooks, dependency errors, and pull request feedback. Route work to specialized skills and agents, with human review checkpoints, diagnosis-only options, and a handoff record.
Route requests to commit, push, or save Git changes through a dedicated /commit skill. The workflow removes Claude attribution, records attempted approaches in reasoning.md, clears build attempts, and pushes changes when requested.
Build Python agents with Agentica SDK using function decorators or agent instances. Examples cover conversation memory, external tools, MCP connections for tool access, multi-agent coordination, typed results, and logging to help implement and debug agent behavior.
Write clearer instructions for Agentica and REPL agents by specifying roles, file operations, and output formats. Includes templates for dividing research, coordinating specialists, reviewing solutions, and passing results between agents through files.
Find and replace code using syntax patterns with ast-grep. Search for function calls, imports, and definitions while excluding matches inside strings or comments, then preview proposed replacements before applying changes across selected files or directories.
Inspect Claude Code execution traces in Braintrust, including tool calls and token usage. The guide explains event hooks, troubleshooting missing records, and matching separate parent and sub-agent traces using timestamps and metadata.
Find unused functions through static analysis, which examines code without running it. Scan a project or directory, choose a language, and exclude designated entry points to get a list of functions with file locations.
Guide implementations that pass session and trace identifiers explicitly between processes and asynchronous steps. It covers storing IDs, using full UUIDs, and separating identifier types to prevent results from being attributed to the wrong session.
Standardize Python scripts that call MCP tools, which connect an agent to external capabilities. The rules cover command-line parameters, usage documentation, tool names, error messages, and output that Claude can process.
Search GitHub code, repositories, issues, and pull requests through MCP, a connection to external tools. Use GitHub search syntax and owner or repository filters to find implementation examples, bug reports, and project information.
Design software that keeps working with reduced features when an optional service is unavailable. The skill provides patterns for early availability checks, cached results, and messages explaining what is missing, which features are affected, and how to restore them.
Discover the tools, workflows, specialist agents, and hooks available in a configured workspace. Interactive guidance and topic-specific commands help you choose relevant capabilities, inspect usage examples, and understand customization options.
Provides a reference for building Claude Code hooks, scripts or model checks triggered by events. Covers input and output formats, registration, testing, and debugging for tasks such as blocking tool calls, adding context, and checking tests before stopping.
Design fallback and repair paths to be idempotent, meaning repeated runs leave the same result. The skill outlines checks before writing, atomic file updates, and reconciliation practices to avoid overwrite conflicts and repeated repair loops.
Coordinate background agents through saved files, keeping lengthy work transcripts out of the main conversation. It defines how to pass results between agents, monitor progress, and verify completed work with tests.
Retrieve web page content from a URL or search query through Firecrawl MCP. Export it as Markdown, HTML, or plain text, with support for extracting structured data from pages.
Consult Agentica APIs when building workflows where multiple AI agents cooperate. This reference covers eleven coordination patterns, agent tracking, shared state, handoffs, and ways to combine results, with pointers to detailed API specifications.
Carry out technical plans in verified stages, with automated checks followed by human review. The skill defines task delegation, progress tracking, and persistent handoff records so work can resume with the context needed for the next task.
Carry out approved technical plans in phases, checking success criteria and recording progress as work proceeds. Larger plans can use separate agents with saved handoff notes, while manual verification keeps users involved at phase boundaries.
Turn math requests into specific command-line instructions using pattern-based routing. It covers calculus, equations, unit conversion, geometry, logical proofs, and plotting, returning a command for the agent to execute with the corresponding math script.
Implement one task from a larger development plan using test-driven development, where tests precede code. Record changes, decisions, test results, and blockers in a handoff document so the next agent can continue with the necessary context.
Guide agents to index files as soon as they are created, making them searchable for subsequent tasks. The skill describes write-triggered indexing and single-file updates instead of waiting for session-end batches or scheduled jobs.
Choose LLM settings for theorem proving, code generation, and creative exploration. It provides examples for output limits and response variety, plus proof planning prompts and a method for generating multiple proof attempts and selecting the best candidate.
Organize implementation work across research, planning, validation, implementation, and review agents. Each stage passes files to the next, while implementation starts with failing tests and the main conversation coordinates the workflow.
Is this your account? Link GitHub to bring it onto your creator page (Japanese)