npx skills add majiayu000/claude-skill-registry --skill riding-codebase-0xhoneyjar-loa-beauvoir --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
The skill directs an agent to ride through a codebase to produce evidence-grounded Loa artifacts. It emphasizes verifying code against claims with file:line citations, flagging issues rather than fixing, and ensuring the grimoire mirrors the code. It outlines a structured workflow from preflight checks, integrity verification, and context discovery to interactive interviews and generating verification claims, all while logging progress to a trajectory file.
How it works
- Preflight & Mount Verification: checks for a mounted Loa version file, then performs system integrity checks against a checksums file, blocking Ride unless integrity is verified or overridden with --force-restore. It also detects the execution context (framework repo vs. target repo).
- Target Resolution (framework repo only): prompts for a codebase target when running from the Loa framework repository.
- Initialize Ride Trajectory: records a preflight entry in a trajectory JSONL file.
- Codebase Probing and Loading Plan: runs a probe to determine total files, lines, and estimated tokens; selects a loading strategy (full, prioritized, or excerpts) based on size; generates a loading plan and logs the probe to trajectory.
- Phase 1: Interactive Context Discovery: checks for existing context, prompts to add context files, analyzes any provided context, and conducts a gap-focused interview using a structured topic list.
- Generate Claims to Verify: mandatory step to create grimoires/loa/context/claims-to-verify.md with a markdown table for architecture, domain, tribal knowledge, and WIP status, including verification strategies. Logs the activity to trajectory.
When to use it
- Use when you need to extract or verify truth from a codebase, producing Loa artifacts that reference exact file:line citations. Activate when starting a project analysis to ensure evidence-backed documentation and flag potential issues rather than attempting fixes.
What it can touch
- Repository files and directories related to the Loa workflow and target codebase, including:
- .loa-version.json
- .claude/checksums.json
- grimoires/loa/a2a/trajectory/
- grimoires/loa/reality/
- grimoires/loa/context/
- target codebase directories detected via TARGET_REPO
- Tools invoked: Read, Grep, Glob, Bash(git *) as per allowed-tools.
Caveats
- Requires the Loa to be mounted and integrity file to exist for blocking behavior; otherwise blocks on integrity or FORCE RESTORE are explicit in the script.
- The workflow emphasizes flagging issues and evidence-backed claims; it does not perform code fixes.
- Output includes prompts and interactive questions that depend on user responses (e.g., target repo selection, context file prompts).
# Riding Through the Codebase You are analyzing an existing codebase to generate evidence-grounded Loa artifacts following the v0.6.0 Enterprise-Grade Managed Scaffolding model. > *"The Loa rides through the code, channeling truth into the grimoire."* ## Core Principles ``` CODE IS TRUTH → Loa channels CODE → Grimoire reflects REALITY ``` 1. **Never trust documentation** - Verify everything against code 2. **Flag, don't fix** - Dead code/issues flagged for human decision 3. **Evidence required** - Every claim needs `file:line` citation 4. **Target repo awareness** - Grimoire lives WITH the code it documents --- ## Phase 0: Preflight & Mount Verification ### 0.1 Verify Loa is Mounted ```bash if [[ ! -f ".loa-version.json" ]]; then echo "❌ Loa not mounted on this repository" echo "" echo "The Loa must mount before it can ride." echo "Run '/mount' first, or:" echo " curl -fsSL https://raw.githubusercontent.com/0xHoneyJar/loa/main/.claude/scripts/mount-loa.sh | bash" exit 1 fi VERSION=$(jq -r '.framework_version' .loa-version.json) echo "✓ Loa mounted (v$VERSION)" ``` ### 0.2 System Zone Integrity Check (BLOCKING) Before the Loa can ride, verify the System Zone hasn't been tampered wit
- Core Principles
- Phase 0: Preflight & Mount Verification
- 0.1 Verify Loa is Mounted
- 0.2 System Zone Integrity Check (BLOCKING)
- 0.3 Detect Execution Context
- 0.4 Target Resolution (Framework Repo Only)
- 0.5 Initialize Ride Trajectory
- Attention Budget
- Token Thresholds
- Clearing Triggers for Codebase Riding
- 4-Step Clearing
- RLM Pattern Alignment
- Semantic Decay Stages
- Phase 0.5: Codebase Probing (RLM Pattern)
if [[ ! -f ".loa-version.json" ]]; then echo "❌ Loa not mounted on this repository" echo "" echo "The Loa must mount before it can ride." echo "Run '/mount' first, or:" echo " curl -fsSL https://raw.githubusercontent.com/0xHoneyJar/loa/main/.claude/scripts/mount-loa.sh | bash" exit 1 fi echo "✓ Loa mounted (v$VERSION)" if [[ ! -f "$CHECKSUMS_FILE" ]]; then
What does the ride skill do?
Analyze codebase to extract reality into Loa artifacts
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill riding-codebase-0xhoneyjar-loa-beauvoir --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.