[Track A] Feedback: Systemic Insecure Code Generation

[rocklambros]

Sprint 2 Feedback — Systemic Insecure Code Generation

This issue collects qualitative feedback on the Systemic Insecure Code Generation new candidate entry. Submit your Importance, Clarity, and Distinctness scores via the Sprint 2 Google Form (single ballot covering both tracks).

• Vote: https://docs.google.com/forms/d/e/1FAIpQLSebmFQyJPOMuw1IorHCJ2FdW98ZT2oLNBuVgYqpPoayW9v6ww/viewform — page through to the Track A section, then to Systemic Insecure Code Generation
• Read the draft: https://github.com/GenAI-Security-Project/GenAI-LLM-Top10/blob/main/2026/new_entry_candidates/systemic-insecure-code-generation.md
• Voting closes: May 18, 2026 at 23:59 UTC
• Code of Conduct: https://owasp.org/www-policy/operational/code-of-conduct

Scoring rubric

• Importance (1–5): How critical is this risk to LLM application security in 2026?
• Clarity (1–5): How clearly does the candidate describe the risk and mitigations?
• Distinctness (1–5): Is this materially different from existing entries (LLM01–LLM10), or could it be merged into one of them?

Why Distinctness matters

Sprint 3 cuts 8 candidates down to 5. A high-Importance candidate that overlaps heavily with an existing entry is a merge target, not a standalone winner. Distinctness scores let the working group make defensible cut, keep, or merge calls.

How to comment

Reply to this issue with feedback specific to Systemic Insecure Code Generation. Use one comment per topic so the working group can triage cleanly. Format we like:

Section: <which section of the candidate>
Issue: <what is wrong, missing, unclear, or duplicative>
Proposed change: <your suggested fix, addition, or merge target>
Overlaps with: <existing entry LLM01–LLM10 if applicable>
Evidence: <CVE, paper, incident, or research note ID if applicable>

Out of scope here

• Cross-cutting comments → use Discussions
• Comments on existing entries (LLM01–LLM10) → comment on the relevant Track B issue
• Process or governance questions → comment in Discussions

Triage

Working group entry leads classify each comment within 48 hours as:

• actionable — feeds Sprint 3 revision or merge decision
• clarification — requires reply, not action
• out-of-scope — closed with redirect

Sprint 3 begins May 18 with this issue's triage as input.

[nagasatish007]

Section: Prevention and Mitigation Strategies

Issue: The mitigations correctly identify SAST gating and secure-coding
constraints, but there's a gap between "full SAST scan" (heavy, CI-only)
and "no check at all" (current state for most inline code generation).
A lightweight governance-layer check at generation time could catch the
highest-signal patterns before code ever reaches a repo.

Proposed change: Add a mitigation for lightweight output-side pattern
detection at the governance layer:

• Detect string interpolation in SQL/command contexts (the README: update Slack channel to #team-genai-top-10-llm #1 failure mode
  per the entry's own data)
• Detect hardcoded credential patterns in generated code (API keys,
  passwords, connection strings)
• Detect missing parameterization in database query construction
• Flag but don't block (REPORT_ONLY mode) to avoid disrupting developer flow

This is NOT a replacement for SAST — it's a fast, deterministic first-pass
that catches the most common and dangerous patterns (CWE-89, CWE-798) at
generation time rather than waiting for CI. Think of it as the governance
equivalent of a spell-checker vs a full editor review.

The entry notes "45-62% of code generation tasks produce vulnerabilities" —
even catching the top 2 patterns (SQL injection + hardcoded credentials) at
the governance layer would meaningfully reduce the volume reaching SAST.

Overlaps with: LLM02 (Sensitive Information Disclosure) for the credential
embedding angle — secret detection in outputs is already a known mitigation
there.

Evidence: The entry's own Scenario #1 (string interpolation SQL) and
Scenario #2 (hardcoded Redis password) are both detectable with simple
regex patterns — no ML required.