By anvanvan
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
Execute plan with fresh subagent per task and code review
Create detailed implementation plan with bite-sized tasks
Interactive design refinement using Socratic method
Execute plan in parallel session with gradual batched checkpoints
Create context handoff for new Claude session - extracts issue, files, commits, and analysis
Use this agent when a major project step has been completed and needs to be reviewed against the original plan and coding standards. Concurrent-safe: uses commit-SHA-list approach to avoid conflicts with parallel work. Examples: <example>Context: The user is creating a code-review agent that should be called after a logical chunk of code is written. user: "I've finished implementing the user authentication system as outlined in step 3 of our plan" assistant: "Great work! Now let me use the code-reviewer agent to review the implementation against our plan and coding standards" <commentary>Since a major project step has been completed, use the code-reviewer agent to validate the work against the plan and identify any issues.</commentary></example> <example>Context: User has completed a significant feature implementation. user: "The API endpoints for the task management system are now complete - that covers step 2 from our architecture document" assistant: "Excellent! Let me have the code-reviewer agent examine this implementation to ensure it aligns with our plan and follows best practices" <commentary>A numbered step from the planning document has been completed, so the code-reviewer agent should review the work.</commentary></example>
Expert web research agent (flows plugin) - finds current documentation, API info, best practices, and technical solutions. Automatically triggered by research keywords or invoke manually with @agent-web-researcher.
Use when creating or developing, before writing code or implementation plans - refines rough ideas into fully-formed designs through collaborative questioning, alternative exploration, and incremental validation. Don't use during clear 'mechanical' processes
Use before refactoring legacy code without tests - documents actual behavior as safety net, capturing current behavior including bugs to prevent regressions during refactoring
Systematically research and understand unfamiliar codebases using parallel Explore agents. Use when encountering unfamiliar code, new projects, or undocumented systems.
Use when tests have race conditions, timing dependencies, or inconsistent pass/fail behavior - replaces arbitrary timeouts with condition polling to wait for actual state changes, eliminating flaky tests from timing guesses
Use when invalid data causes failures deep in execution, requiring validation at multiple system layers - validates at every layer data passes through to make bugs structurally impossible
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
This fork uses git stage-lines and git unstage-lines for granular staging control, which is essential for parallel agent workflows. These are custom git aliases that require setup.
Run the automated setup script:
bash ~/tools/flows/setup.sh
This will install patchutils (if needed) and configure the git aliases automatically.
If you prefer to set up manually:
brew install patchutils
This provides the filterdiff command used by the git aliases.
Add these aliases to your global git configuration:
git config --global alias.stage-lines '!f() {
# Handle parameter patterns
if [ -z "$2" ]; then
file="$1"
lines=""
else
lines="$1"
file="$2"
fi
# Check if file is tracked
if git ls-files --error-unmatch "$file" 2>/dev/null; then
# Tracked file: require line ranges
if [ -z "$lines" ]; then
echo "Error: Please specify line ranges (e.g., 1-5,8,10-12)" >&2
echo "Usage: git stage-lines <lines> <file>" >&2
echo " git stage-lines <untracked-file>" >&2
return 1
fi
# Use filterdiff for line-level staging
git diff "$file" | filterdiff --lines="$lines" | git apply --cached --unidiff-zero
else
# Untracked file: stage entire file
git add "$file"
fi
}; f'
git config --global alias.unstage-lines '!f() {
# Handle parameter patterns
if [ -z "$2" ]; then
file="$1"
lines=""
else
lines="$1"
file="$2"
fi
# Check if file is newly added
if git diff --cached --diff-filter=A --name-only | grep -qx "$file"; then
# Newly added file: remove from index, make untracked
git rm --cached "$file"
else
# Partial staged changes: require line ranges
if [ -z "$lines" ]; then
echo "Error: Please specify line ranges (e.g., 1-5,8,10-12)" >&2
echo "Usage: git unstage-lines <lines> <file>" >&2
echo " git unstage-lines <newly-added-file>" >&2
return 1
fi
# Use filterdiff reverse for line-level unstaging
git diff --cached "$file" | filterdiff --lines="$lines" | git apply --cached --unidiff-zero --reverse
fi
}; f'
Check that aliases are configured:
git config --global --get-regexp "alias\.(stage-lines|unstage-lines)"
You should see output showing the configured aliases (formatting may vary).
The commands automatically detect file state and apply the appropriate staging behavior:
For tracked files with changes (requires line ranges):
# Stage specific lines from a tracked file
git stage-lines 10-25 path/to/file.py
# Unstage specific lines from staged changes
git unstage-lines 15-20 path/to/file.py
For untracked or newly-added files (line ranges optional):
# Stage an entire untracked file
git stage-lines new-file.js
# Unstage a newly-added file (makes it untracked again)
git unstage-lines new-file.js
The commands detect whether a file is tracked or untracked and automatically use the appropriate staging method. For tracked files, line ranges are required to enable granular control. For untracked files, the entire file is staged.
This allows multiple parallel agents to work on the same files and stage only their specific changes without conflicts.
Flows includes an intelligent web research agent that automatically activates when you need current external information.
The agent activates when you mention research keywords:
Explicitly invoke the agent in any context:
@agent-web-researcher find the latest Next.js 14 app router patterns
The web-researcher agent:
The agent is integrated throughout flows skills with strategic invocation points in debugging, brainstorming, planning, and skill-creation workflows.
Flows includes a concurrent-safe code review system designed for environments where multiple agents or humans work simultaneously.
npx claudepluginhub anvanvan/flowsHarness-native ECC plugin for engineering teams - 67 agents, 278 skills, 94 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
Superpowers Plus core skills library for Claude Code: planning, execution routing, TDD, debugging, and collaboration workflows
Evidence-gated AI coding workflow: scan → analyze → plan → TDD → execute → fix → verify → review, powered by Codebase Memory MCP >= 0.9.0 with optional Serena LSP intelligence. Includes blast-radius planning, test/cycle gates, independent review, and Windows Git Bash hook auto-resolution.
v9.52.0 - Reliability wave: tangle contextual review correction loop with hard round ceiling, progress-supervised review rounds (per-agent stall watch, descendant-tree kills), council diversity and agy pin fixes, marketplace generator source-of-truth fix, provider troubleshooting runbook and cost-expectations docs. Run /octo:setup.
Claude harness - A harness for solo developers (Vibecoders) to handle full-cycle contract development.