Plugin Claude Code
Agent Skills and plugins for Claude Code by astamvalley.
Plugin Claude Code
Agent Skills and plugins for Claude Code by astamvalley.
Plugin Claude Code
Crucible — full SDLC coverage from planning to implementation and review. Dual Solo/Team mode, adversarial review, and natural language entry. Commands: cast, mold, forge, smelt, assay, security, challenge, debug, auto, stats, status, result, help.
Command
A code-review command that examines files, changes, or supplied code from architectural, logic, and security angles, then combines the findings.
Command
A natural-language entry point for using CRB, which selects and runs a suitable CRB command from a description of the situation. CRB is not explained further in the input.
Command
A planning workflow that uses analysis from several angles and brings the results toward agreement. It is intended for game plans, app designs, architecture decisions, and product direction, and is called with /crb:cast or /crb:cast --auto.
Command
An adversarial review command for design or planning work. It takes an existing cast output file or a topic and produces objections and risks.
Command
A command for analysing error messages, stack traces, or described software symptoms to identify likely causes, affected areas, and possible fixes.
Command
A command for implementing code and having multiple AI reviewers check it. It can also use TDD, a method where tests are written before the code.
Command
A command that explains the available commands and their usage, either with a general summary or details for a named command.
Command
A technical design command that turns a feature name or description into a plan covering the system structure, data, components, and edge cases.
Command
A result command that finds the output file named in the last entry of run-log.jsonl and checks whether that file exists.
Command
A security-checking command that examines code, features, or designs for vulnerabilities using three areas: the OWASP Top 10 list, login and permission flows, and data exposure.
Command
A setup command for the Claude Foundation project. It checks the recommended RTK installation and optionally checks agent teams and the Codex command-line tool.
Command
A workflow that turns a written specification into code through design, implementation, and review stages.
Command
A command that reads a JSONL run log, a text file with one JSON record per line, and analyzes usage patterns.
Command
A command that reads .crb/runs/run-log.jsonl, a file containing CRB run records, and displays the execution history.
Skill Claude CodeCodex
Nástroj, ktorý v koreňovom priečinku projektu automaticky rozpozná používaný programovací jazyk a framework. Výsledok uloží do súboru .crb/project-context.md na opakované použitie.
Skill Claude CodeCodex
A command for exploring a codebase or system by detecting the model, choosing a perspective, and running exploration tasks in parallel.
Skill Claude CodeCodex
A skill for configuring crb-forge, including agent setup, normal mode, and TDD mode. TDD, or test-driven development, means writing tests before or alongside the code they check.
Skill Claude CodeCodex
A multi-step review process with separate modes, parallel cross-checking, agreement-building, and user confirmation when automatic approval is not enabled.
Skill Claude CodeCodex
Skill "crb-output" from astamvalley/claude-foundation, covering crb-output, 디렉토리 구조, 세션 id 생성, run-log.jsonl 형식 and 출력 파일 경로.
Skill Claude CodeCodex
A checkpoint and resume helper for long-running commands, including deep runs of the named tools. The description is limited to detecting checkpoints and continuing interrupted runs.
Skill Claude CodeCodex
An internal decision skill for choosing between solo work and Agent Teams, where multiple agents collaborate on a task.
Skill Claude CodeCodex
A tool that examines changed code, creates a Git commit message following the repository's convention, and runs the commit. Git is the version-control system used to record code changes.