Skills from clawhub.ai/atlaspa

This skill scans a workspace for prompt-injection patterns and analyzes content boundaries. It scans the entire workspace by default (`~/.openclaw/workspace`) and includes commands that run `python3 scripts/bastion.py` and clone `https://github.com/AtlasPA/openclaw-bastion.git`.

Tamper-evident audit trail for agent sessions. Hash-chained logs of workspace changes with cryptographic verification. Know exactly what changed, when, and detect if anyone altered the record. Free alert layer — upgrade to openclaw-ledger-pro for freeze, forensics, and chain restoration.

This skill is a prompt-injection defense tool that scans a workspace, modifies or quarantines files, deploys canaries, and generates runtime hooks. It executes shell commands (`python3 scripts/bastion.py`, `git clone https://github.com/AtlasPA/openclaw-bastion-pro.git`), and writes to agent identity files (`SOUL.md`, `AGENTS.md`, `.quarantine/bastion`, `.bastion`).

This skill scans an agent workspace for outbound network calls, suspicious domains, and data-exfiltration patterns. It includes explicit instructions to run `python3 {baseDir}/scripts/egress.py`, a `git clone https://github.com/AtlasPA/openclaw-egress.git` install step, and the code reads the `OPENCLAW_WORKSPACE` env var.

This skill audits agent workspaces for command, network, and data-handling policy compliance and generates audit-ready reports. It includes example shell commands to run the tool (e.g., `python3 scripts/marshal.py policy --init`), which instructs local script execution but shows no external secret exfiltration.

This skill is a network DLP tool that scans an OpenClaw workspace for outbound URLs and can block or quarantine skills. It modifies workspace files and skill directories (`.egress-allowlist.json`, `skills`, `.quarantined-`) and exposes CLI commands like `python3 scripts/egress.py` that perform those actions.

OpenClaw Ledger Pro provides hash-chained workspace auditing, forensic analysis, restoration, and automated protection sweeps. It includes `python3 scripts/ledger.py` commands that read/write `.ledger` artifacts, references `OPENCLAW_WORKSPACE`, and shows a `git clone https://github.com/AtlasPA/openclaw-ledger-pro.git` example.

High-risk orchestrator that runs local skill scripts and external installers against agent workspaces. It executes shell commands (`subprocess.run`, `python3 scripts/security.py`) and performs network installs (`clawhub`, `git clone https://github.com/AtlasPA/openclaw-security.git`) using `--workspace` paths like `/path/to/workspace` and `~/.openclaw/workspace`.

Provides a local supply-chain scanner and automated countermeasures (scan, quarantine, SBOM, monitor) via `scripts/sentinel.py` operating on `~/.openclaw/workspace`. It instructs executing shell commands (`python3 scripts/sentinel.py ...`), cloning `https://github.com/AtlasPA/openclaw-sentinel-pro.git`, and adding a `SessionStart` hook to run protection at startup.

This skill audits installed agent skills for network, subprocess, file I/O, serialization, and environment usage. Documentation includes commands that run `python3 scripts/arbiter.py`, an installation `git clone https://github.com/AtlasPA/openclaw-arbiter.git`, and the code reads `OPENCLAW_WORKSPACE`.

This skill scans installed agent skills for supply-chain risks and stores a local threat database and scan history in the workspace `./.sentinel` (`threats.json`, `history.json`). No security-relevant behaviors detected.

This skill is a compliance and enforcement suite for OpenClaw that audits skills, generates runtime hooks, and can auto-quarantine non-compliant skills. It instructs running shell commands like `python3 scripts/marshal.py protect`, renames skill directories with the `.quarantined-` prefix, and suggests modifying ` .claude/settings.json`.

This skill provides a local secret-scanning and countermeasure tool that scans, redacts, quarantines, and modifies workspace files and policies (e.g., `scripts/sentry.py`, `.gitignore`, `.sentry-policy.json`). It instructs executing `python3 scripts/sentry.py protect`, reads `OPENCLAW_WORKSPACE`, creates `.bak` backups, and moves files into `.quarantine/sentry/`, requiring broad filesystem modification and command execution.

This skill scans agent workspaces for leaked credentials and high-risk files. It reads workspace files (including `.env`) and the `OPENCLAW_WORKSPACE` env var, and its documentation includes commands like `python3 scripts/sentry.py scan` that instruct running the scanner.

This skill computes SHA-256 signatures for installed skills and manages a trust manifest (`.signet/manifest.json`) via CLI commands. It includes runnable shell commands (e.g., `python3 scripts/signet.py sign`, `git clone https://github.com/AtlasPA/openclaw-signet.git`), reads `OPENCLAW_WORKSPACE`, and writes to `~/.openclaw/workspace`.

This skill provides local integrity management for an OpenClaw workspace with commands like `sign`, `verify`, `protect`, `quarantine`, `snapshot`, and `restore`. It includes executable hooks that run `python3 scripts/signet.py` (e.g., a `SessionStart` hook) and installation instructions using `git clone https://github.com/AtlasPA/openclaw-signet-pro.git`, and reads `OPENCLAW_WORKSPACE`.

OpenClaw Triage inventories an agent workspace, builds timelines, scans files for credential/exfil patterns, and preserves evidence locally via `python3 scripts/triage.py` with output under `.triage/evidence-{timestamp}/`. It reads ` .integrity/manifest.json`, `.ledger/chain.jsonl`, `.signet/manifest.json`, and `.sentinel/threats.json`; these behaviors are purpose-aligned and low risk.

Provides a local incident-response suite for agent workspaces: investigate, contain, remediate, collect evidence, harden, and run playbooks. It scans and copies workspace files, disables `.claude` hooks, moves skills to `.triage/quarantine`, and runs local remediation scripts via `subprocess.run` while honoring `OPENCLAW_WORKSPACE`.

This skill audits credential exposure across a workspace, scanning `.env` files, shell history, git configs, logs, Docker files, and building an inventory. It reads sensitive local files (e.g., `~/.bash_history`, `~/.gitconfig`, `~/.openclaw/workspace`) and includes install/run commands (`git clone https://github.com/AtlasPA/openclaw-vault.git`, `python3 scripts/vault.py ...`).

This skill scans workspaces for credentials and automates remediation (permission fixes, quarantine, rotation, git-history scans). It runs local commands (e.g., `python3 scripts/vault.py protect`), executes subprocesses (`git`, `icacls`), reads home files like `~/.bash_history`, and suggests changing agent config (`.claude/settings.json`).

Monitors workspace file integrity and scans for prompt injection patterns in agent files. Contains instructions to run `python3 scripts/integrity.py` locally and to clone `https://github.com/AtlasPA/openclaw-warden.git` via `git`.

This skill scans an OpenClaw agent workspace for integrity and prompt-injection issues and provides automated countermeasures (snapshot restore, git rollback, skill quarantine, and `protect`). It includes explicit execution instructions (e.g., `python3 scripts/integrity.py protect`, `git clone https://github.com/AtlasPA/openclaw-warden-pro.git`) and modifies workspace files such as `SOUL.md` and `skills/*/`.