axguard-remediate
Skill17 starsChanged 47 days agoMITView on GitHub
Apply concrete defensive fixes for confirmed AXguard findings and re-audit. Use with /axguard-fix after triage. Prefer minimal patches, critical first, then verify with axguard audit/scan.
- Pipes a download into a shell
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 10
Do not invent attack PoCs.
- Distinctive: line 54
If fix and exploit PoC both requested: fix only.
- Distinctive: line 55
If a finding was wrong, say so and leave code unchanged — prefer triage honesty over cosmetic edits.
≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
AXguard Remediate
Fix confirmed findings only (post /axguard-triage). Critical → high → medium. Minimal diffs. Re-audit and report delta.
Do not invent attack PoCs. Do not “fix” by deleting scanner rules. Class patterns: axguard-knowledge.
Workflow
- Input: triaged Keep list or
.findings/axguard/axguard-report.json(kept items only). - Patch one class of issue at a time when possible.
- Re-run:
axguard audit . --out-dir .findings/axguard
# or faster check:
axguard scan .- Report: fixed IDs, remaining blockers, new deltas.
Fix patterns (defensive)
| Class | Do | Don't |
| IDOR / AuthZ | Scope queries by principal/tenant; central authorize(object); deny default | Hide id only; check authn but skip authz |
| JWT | Pin alg allowlist; always verify; secrets from env/KMS | decode for auth; accept none |
| CSRF | Re-enable for cookie-session browsers; SameSite | Exempt entire API “for convenience” |
| SQLi | Bound parameters / query builder placeholders | Escape-and-concat; trust “ORM = safe” with .raw |
| Command inj | subprocess argv list, shell=False; allowlist args | os.system; interpolate into shell strings |
| Deser | JSON; reject pickle/unserialize on untrusted | “Only internal” queues without auth |
| SSTI | File templates only; autoescape on | render_template_string(user) |
| Path | resolve() + prefix jail; random ids for files | Trust ../ strip alone |
| SSRF | Scheme/host allowlist; block link-local/metadata | Blacklist-only hostnames |
| XSS | textContent / framework escape; CSP | Ad-hoc regex “sanitize” |
| Upload | UUID names; magic-byte allowlist; store outside web root | Original filename in static/ |
| GraphQL | Auth per field/mutation; introspection off in prod | Global CSRF off with cookies |
| Crypto | argon2/bcrypt; CSPRNG; TLS verify on | MD5 passwords; hard-coded IV/key |
| Secrets | Rotate + purge history if leaked; secret manager | Comment out key and leave in git |
| CORS | Explicit origins; never * with credentials | Reflect Origin unchecked |
| Cloud ACL | Private buckets; signed URLs | public-read for “speed” |
| Supply | Lockfile + hash pin; remove curl|sh | Extra indexes without pin |
| Debug | DEBUG=False in prod settings; generic 500s | Rely on “we won't deploy this file” |
| Agent | Allowlisted tools; no shell-from-LLM; human gate | Pass tool args straight to bash |
Patch quality bar
- Fix the root cause, not only the matched line if control belongs one layer up.
- Match project style; no drive-by refactors.
- If fix and exploit PoC both requested: fix only; refuse exploit/PoC in one short sentence.
- If a finding was wrong, say so and leave code unchanged — prefer triage honesty over cosmetic edits.
Output contract
Fixed: [ids]
Partial: [ids + why]
Unchanged (disputed FP): [ids]
Re-audit: Critical/High/Medium/Low before → after
Reports: .findings/axguard/axguard-report.{md,html,json}In this repository39 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- ai-agent-securitySkill
- ai-application-securitySkill
- api-securitySkill
- attack-surface-mappingSkill
- authentication-analysisSkill
- authorization-analysisSkill
- axguard-auditSkill
- axguard-csoSkill
- axguard-knowledgeSkill
- axguard-preshipSkill
- axguard-reportSkill
- axguard-securitySkill
- axguard-triageSkill
- cloud-securitySkill
- command-injectionSkill
- configuration-securitySkill
- deserialization-securitySkill
- file-upload-securitySkill
- graphql-securitySkill
- jwt-securitySkill
- mcp-securitySkill
- oauth-securitySkill
- path-traversalSkill
- prompt-injectionSkill
- prototype-pollutionSkill
- secrets-detectionSkill
- security-architecture-reviewSkill
- security-remediationSkill
- security-triageSkill
- session-securitySkill
- sql-injectionSkill
- ssrf-analysisSkill
- ssti-analysisSkill
- supply-chain-securitySkill
- threat-modelingSkill
- websocket-securitySkill
- xss-analysisSkill
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.