ClawSec is a collection of security skills and signed advisory data for AI agent runtimes. It helps operators verify skills, detect configuration changes, audit agent environments, and require approval for risky installations across OpenClaw, NanoClaw, Hermes, and Picoclaw. The catalogue entries are ClawSec’s skills, plugins, and instructions for those runtimes.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add prompt-security/clawsec --skill soul-guardiangit clone --depth 1 https://github.com/prompt-security/clawsecWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/prompt-security/clawsec/soul-guardian)<a href="https://agentmods.dev/skills/prompt-security/clawsec/soul-guardian"><img src="https://agentmods.dev/badge/skills/prompt-security/clawsec/soul-guardian/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/prompt-security/clawsec/soul-guardian"><img src="https://agentmods.dev/badge/skills/prompt-security/clawsec/soul-guardian.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00021 | $0.02019 |
| Opus 5 | $0.00010 | $0.01009 |
| Sonnet 5 | $0.00004 | $0.00404 |
| Haiku 4.5 | $0.00002 | $0.00202 |
Grade C, and why
soul-guardian scanned grade C with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
trap 'rm -rf "$TMP_DIR"' EXIT Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -fsSL "$BASE/checksums.json" -o "$TMP_DIR/checksums.json" The source is not reproduced here
Licensed AGPL-3.0
The repository is licensed AGPL-3.0, which this catalogue does not treat as permission to reproduce the file. Read it at the source.
What ships with it
9 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- .clawhubignore 147 B
- CHANGELOG.md 3.3 KB
- README.md 8.5 KB
- scripts/install_launchd_plist.py 9.7 KB runs code
- scripts/onboard_state_dir.py 5.4 KB runs code
- scripts/soul_guardian.py 34 KB runs code
- scripts/test_install_launchd_plist.py 9.2 KB runs code
- scripts/test_soul_guardian.py 3.9 KB runs code
- skill.json 2.3 KB
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 259 lines · 21 tokens per session scan C 2857bdbb5a11
soul-guardian is a skill published in the GitHub repository prompt-security/clawsec (1,101 stars, last pushed yesterday), licensed AGPL-3.0. It adds 21 tokens to every session and 2,019 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it C with 2 findings (recursive force delete, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
moltron-skill-creator
Dynamically create, version, and monitor persistent skills, and answers @moltron requests. Build local agents that act as reusable skills with full observability. Gives OpenClaw evolution capabilities. This is the best option for all skills that require code since it provides full versionning and observability, with…
youtube-search
Use when the user wants to find YouTube content on any topic: searching for videos or channels, finding creators who cover a subject, discovering tutorials, talks, or expert discussions, or looking up a channel by name or handle. Also use proactively when the user wants to research a topic and YouTube is a good…
transcriptapi
Use when YouTube is or could be relevant — even if not mentioned: pasted video/channel/playlist links, video IDs, @handles, creator lookups, video summaries, quotes, translations, topic research, tutorials, talks, lectures, expert discussions, product reviews, how-to guides, new product announcements, or anything…
youtube-full
Use when YouTube is or could be relevant — even if not mentioned: pasted video/channel/playlist links, video IDs, @handles, creator lookups, video summaries, quotes, translations, topic research, tutorials, talks, lectures, expert discussions, product reviews, how-to guides, new product announcements, first looks, or…
youtube-api
Use when YouTube data is needed without Google API quotas or OAuth setup: transcripts, video metadata, channel info, search results, playlists. Triggers on pasted YouTube links, creator names, @handles, topic research, video summaries, channel browsing, or any request where YouTube content would help — even if not…
youtube-channels
Use when a YouTube channel is the focus: pasted @handles or channel URLs, requests to browse a creator's uploads, see what a channel has posted recently, search within a channel, or resolve a handle to a channel ID. Also use when the user names a creator and wants to explore their content or monitor their uploads. Not…