Repository navigation
fix: resolve idna CVE-2024-3651 false negative via Python-scoped reachability gate - #406
heatherzh01 wants to merge 4 commits into
Conversation
|
/test-heavy |
|
/test vulnerability-analysis-on-pr |
|
/test-heavy |
1 similar comment
|
/test-heavy |
Automated multi-agent review — idna CVE-2024-3651 Python-scoped gateReviewed across five dimensions (correctness, tests, silent failures, comments, general quality). The two most severe findings were verified by hand against the branch. Headline: as written, the fix does not fire — the new gate is currently dead code. 🔴 CRITICAL1. The fix never fires — 2. Marker substring-match runs over source-contaminated tool output → false-positive 3. 🟠 IMPORTANT4. The 5. "Keep markers in sync" is a manual coupling that fails silently toward false negative. 6. Keyword-header format coupling fails silently toward false negative. 7. Inaccurate comment about the non-Python path. 8. CVE/ticket references in source comments (project convention: describe behavior, not task origin). 9. Test coverage: four bug-fix functions ship with no tests; no plumbing test. 10. CCA tool exceptions are swallowed and indistinguishable from "no analysis". 🟡 Suggestions
|
|
Caution There are some errors in your PipelineRun template.
|
…python-scoped idna fix
This comment was marked as duplicate.
This comment was marked as duplicate.
|
/test-heavy |
|
/test vulnerability-analysis-on-pr |
1 similar comment
|
/test vulnerability-analysis-on-pr |
|
/test-heavy |
|
Question on the The Rule 8 prompt rewrite should prevent the model from generating However, the guard currently does a hard reject on all Two questions:
|
Summary
Fixes a false negative where idna CVE-2024-3651 (quadratic-complexity DoS in
idna.encode) was scored not exploitable even though the sample applicationcalls
idna.encode()directly on user-provided input in its Flask handler.The fix is scoped entirely to the Python ecosystem — Go, C, and Java verdicts
are byte-for-byte unchanged from baseline.
Root cause
idna.encode),the reachability agent sent the package name as the function name
(
"idna,idna"). Call Chain Analyzer looked upidna()instead ofencode(),returned "function not found", and that failure was recorded as a genuine
"not reachable" verdict.
real negative verdict, tripping the hard reachability gate — so even with a
direct
idna.encode()call site in application source, the verdict was forcedto not-exploitable.
Changes
react_internals.pymodule.functionsymbol uses the function's own name (encode), never the package name; neverfunction_name == package_name.transitive_code_search.py_validate_query_formatrejects the degeneratepackage==functionquery with a retry hint; newcca_output_indicates_no_analysis()/CCA_NO_ANALYSIS_MARKERSdistinguish a failed/blocked CCA result from a real verdict.reachability_agent.pyFalseis kept out ofcca_results;app_source_callsiteis recorded on a main-application-source hit. Other ecosystems keep baseline behaviour.full_text_search.pykeyword_result_has_app_source_hit()+APP_SOURCE_SECTION_LABEL.cve_agent.pyapp_source_callsitethrough to checklist items.cve_summarize.pyuncertainhedging.Why the gate is Python-scoped
The inference "direct application-source call site ⇒ reachable ⇒ lean
exploitable" is sound for Python top-level library functions
(
idna.encode(user_input)), but it over-reports Go DoS CVEs whose vulnerablepath is library-internal. It is confined to Python via
_app_callsite_gate_applies(items, ecosystem).Scope / explicitly out of scope
reachability work, and the Go-DoS over-report problem — separate efforts.
Risk
Low. All behavioural changes are gated on
ecosystem == "python"; the singleglobal change only blocks a malformed
package==functionquery and returns aretry hint.