You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 44c08a9
Browse filesBrowse the repository at this point in the historyBrowse files
fix: exclude valid OMS signatures from content analysis
OMS bundles necessarily contain long base64-encoded payload, signature, and certificate fields. Generic obfuscated-code checks can misclassify these fields as hidden executable content.
Recognize root-level skill.oms.sig files with the minimal OMS v0.3 DSSE/in-toto structure and inventory them as oms_signature components without sending their contents through static or LLM analyzers.
Preserve normal scanning for malformed, nested, oversized, or unrecognized signature files. Add a pinned real-world fixture, regression coverage, completeness accounting, and documentation.
Recognition is structural only and does not verify signatures, certificates, transparency logs, or signer identity.
Signed-off-by: major-security <[email protected]>
- Live vulnerability lookups via OSV.dev for known CVEs in dependencies
654
-
- Scans all files in the skill
654
+
- Scans all analyzer-eligible files in the skill
655
655
- High recall (catches most issues)
656
656
- Moderate precision (some false positives)
657
657
658
+
A valid, root-level OpenSSF Model Signing signature (`skill.oms.sig`) is retained in the
659
+
component inventory as type `oms_signature`, but excluded from static and LLM content analysis.
660
+
OMS bundles necessarily contain long base64-encoded payload, signature, and certificate fields;
661
+
generic obfuscated-code checks can otherwise misclassify those fields as hidden executable content.
662
+
The recognizer checks the minimal OMS DSSE/in-toto structure; it does not verify the signature,
663
+
certificate chain, transparency-log entry, or signer identity. Invalid or unrecognized signature
664
+
files are scanned normally.
665
+
658
666
### Stage 2: LLM Semantic Analysis (Optional)
659
667
- Evaluates context and intent
660
668
- Filters false positives
@@ -679,7 +687,7 @@ The tool requires outbound HTTPS access to `api.osv.dev` for live vulnerability
679
687
SkillSpector is defense-in-depth, not a sandbox. Know what it does and does not do before relying on it:
680
688
681
689
-**It never executes the scanned skill.** All analysis is static (regex, Python AST, YARA) plus optional LLM evaluation of file *contents* — the skill's code is never run.
682
-
-**LLM analysis sends file contents to the configured provider.** When LLM analysis is enabled (the default), file contents are sent to the active `SKILLSPECTOR_PROVIDER` endpoint. Use `--no-llm` to keep contents local (static analysis only).
690
+
-**LLM analysis sends analyzer-eligible file contents to the configured provider.** When LLM analysis is enabled (the default), file contents are sent to the active `SKILLSPECTOR_PROVIDER` endpoint. Recognized OMS signature files are excluded. Use `--no-llm` to keep contents local (static analysis only).
683
691
-**SC4 sends dependency names to OSV.dev.** The supply-chain check queries [OSV.dev](https://osv.dev) with the package names and versions the skill declares, to look up known CVEs. This is fundamental to the check and runs even with `--no-llm`. It sends dependency coordinates (not file contents), requires no API key, and falls back to a bundled list when OSV.dev is unreachable.
684
692
-**It does not sandbox the host.** SkillSpector flags risky patterns *before* you install a skill; it does not contain or isolate a skill you choose to install anyway.
0 commit comments