autonomous-agent
Before main skill execution, perform guardrail checks.
npx skills add majiayu000/claude-skill-registry --skill autonomous-agent --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
Before main skill execution, perform guardrail checks, including input guardrails, danger level enforcement, PII filtering, injection detection, and trajectory logging. It then operates as an Autonomous Agent Orchestrator that executes autonomous work with compliance, self-auditing, remediation loops, and continuous improvement.
How it works
- Step 1: Read .loa.config.yaml to decide if input guardrails are enabled; if guardrails.input.enabled is false or LOA_GUARDRAILS_ENABLED is false, skip to skill execution.
- Step 2: Run .claude/scripts/danger-level-enforcer.sh --skill autonomous-agent --mode {mode}; in Interactive mode require explicit reason; in Autonomous mode BLOCK unless --allow-high is supplied.
- Step 3: Run .claude/scripts/pii-filter.sh to redact sensitive data before orchestration.
- Step 4: Run .claude/scripts/injection-detect.sh --threshold 0.65; use a lower threshold due to higher impact potential.
- Step 5: Log to grimoires/loa/a2a/trajectory/guardrails-{date}.jsonl; on error, log to trajectory and fail-open (continue to skill).
- Then proceed with the Autonomous Agent Orchestrator flow which includes preflight, discovery, design, and implementation phases with audit, remediation loops, and continuous improvement as per the broader framework.
When to use it
- Triggered when autonomous orchestration of tasks with guardrails is required and the system must enforce high risk gates and self-auditing.
- Used in environments where guardrails, PII protection, and risk scoring are mandatory before autonomous execution.
What it can touch
- Tools explicitly listed: .claude/scripts/danger-level-enforcer.sh, .claude/scripts/pii-filter.sh, .claude/scripts/injection-detect.sh
- Logs to grimoires/loa/a2a/trajectory/guardrails-{date}.jsonl
- Reads .loa.config.yaml and environment LOA_GUARDRAILS_ENABLED
Caveats
- The skill is marked as high danger due to orchestration through other skills; requires --allow-high for autonomous mode to proceed.
- If an error occurs during preflight guardrails, the policy is to log to trajectory and continue (fail-open).
- The summary notes that this orchestrator includes integration references and a fixed guardrail workflow; no outcomes are guaranteed by the skill itself.
<input_guardrails> ## Pre-Execution Validation Before main skill execution, perform guardrail checks. ### Step 1: Check Configuration Read `.loa.config.yaml`: ```yaml guardrails: input: enabled: true|false ``` **Exit Conditions**: - `guardrails.input.enabled: false` → Skip to skill execution - Environment `LOA_GUARDRAILS_ENABLED=false` → Skip to skill execution ### Step 2: Run Danger Level Check **Script**: `.claude/scripts/danger-level-enforcer.sh --skill autonomous-agent --mode {mode}` **CRITICAL**: This is a **high** danger level skill (full orchestration control). | Mode | Behavior | | ----------- | --------------------------------------------- | | Interactive | Require explicit confirmation with reason | | Autonomous | BLOCK (high-risk skill requires --allow-high) | **Note**: The autonomous-agent skill is classified as `high` rather than `critical` because it operates through other skills that have their own guardrails. The orchestrator itself doesn't directly execute dangerous operations. ### Step 3: Run PII Filter **Script**: `.claude/scripts/pii-filter.sh` Detect and redact sensitive data before orchestration begins. Important for multi-phase execution where data flows betw
- Pre-Execution Validation
- Step 1: Check Configuration
- Step 2: Run Danger Level Check
- Step 3: Run PII Filter
- Step 4: Run Injection Detection
- Step 5: Log to Trajectory
- Error Handling
- Issue Integrations
- Resources
- Prime Directive
- Execution Model
- Operator Detection
- Configuration
- Detection Heuristics (when type: auto)
Run cleanup in autonomous mode case $CLEANUP_EXIT in if echo "$CLEANUP_RESULT" | jq -e '.partial_state != null' > /dev/null 2>&1; then if [[ -n "$PARTIAL" ]]; then echo "HALT: Partial archive state detected - manual intervention required" echo "Found: $PARTIAL" exit 1 fi echo "✓ Workspace cleanup complete" echo "HALT: Workspace cleanup security validation failed"
What does the autonomous-agent skill do?
Before main skill execution, perform guardrail checks.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill autonomous-agent --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.