fix(ai-review): neutralize public markdown output

[JSONbored]

Motivation

• The AI review path published LLM-derived text into trusted bot comments while only rejecting forbidden product/private terms, leaving markdown, mentions, links, and control characters intact and enabling public-comment injection/spoofing.
• Neutralize AI-generated output before it reaches the public panel so attacker-controlled prompt injection cannot cause preserved markdown (links, images, mentions, lists) to appear in a trusted automation surface.

Description

• Add a neutralizePublicMarkdown helper that flattens control characters/extra whitespace, defangs mentions and :// sequences, and escapes common Markdown metacharacters.
• Apply the neutralizer inside toPublicSafe after the existing sanitizePublicComment check so only allowed content is published and any remaining Markdown/mentions/URLs are neutralized.
• Update the PR panel text in buildPublicPrIntelligenceComment to document that notes are markdown-neutralized upstream while keeping the final angle-bracket escaping guard.
• Add/adjust unit coverage in test/unit/ai-review.test.ts to assert neutralization of links, image syntax, mentions, and newline/list injection.

Testing

• Ran npm test -- --run test/unit/ai-review.test.ts and all unit tests passed (31/31).
• Ran npm run typecheck (tsc --noEmit) and type checking succeeded.

---

Codex Task

[ghost]

Note

Gittensory Gate skipped

PR closed before full evaluation. No late first comment was created.

Signal	Result	Evidence	Action
Gate result	⚠️ Skipped	#902 is no longer open.	No action.

---

💰 Earn for open-source contributions like this. Gittensor lets GitHub contributors earn for the work they already do — register to start earning →.

Checked by Gittensory, a quiet PR intelligence layer for OSS maintainers.

[superagent-security]

Superagent didn't find any vulnerabilities or security issues in this PR.

[ghost]

gittensory · advisory review

Reviewed 3 changed file(s) — two independent AI reviewers.

Changed files: src/services/ai-review.ts, src/signals/engine.ts, test/unit/ai-review.test.ts

Suggested action: ✅ Safe to merge — both reviewers found no blocking issues.

Reviewer A · gpt-oss-120b — recommends ✅ merge
The PR introduces a markdown neutralization step for public‑safe text, enhancing security by stripping control characters, breaking mentions/links, and escaping markdown syntax, and adds a corresponding unit test.

Suggestions

• Document the purpose of neutralizePublicMarkdown in a comment near its definition for future maintainers.
• Consider adding a test case that verifies control characters and excessive whitespace are correctly normalized.
• Review any UI components that render toPublicSafe output to ensure the escaped markdown still displays as intended.

Worth double-checking

• Escaping all markdown characters may affect legitimate formatting in public notes; verify that this is the desired behavior across the UI.
• Zero‑width spaces inserted into URLs/mentions could impact downstream tooling that parses such strings.

Reviewer B · mistral-small-3.1-24b-instruct — recommends ✅ merge
This PR adds a markdown neutralization function to ensure public-facing text is safe and neutralized. The changes are well-targeted and include corresponding tests.

Suggestions

• Consider adding comments to explain the purpose of each regex replacement in the neutralizePublicMarkdown function for better maintainability.
• Ensure that the neutralizePublicMarkdown function handles all edge cases, such as nested or escaped special characters.

Worth double-checking

• Verify that the neutralizePublicMarkdown function correctly handles all potential markdown injection vectors.
• Ensure that the new function does not introduce performance bottlenecks, especially for large text inputs.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.71%. Comparing base (5d9d73a) to head (d9dcc3f).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #902   +/-   ##
=======================================
  Coverage   96.71%   96.71%           
=======================================
  Files         108      108           
  Lines       14641    14642    +1     
  Branches     5298     5298           
=======================================
+ Hits        14160    14161    +1     
  Misses        102      102           
  Partials      379      379           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.