feat: improve error resilience, add provider presets, boxed report outtput, and enhanced README by tejasprasad2008-afk · Pull Request #6 · ni5arga/deanonymizer

tejasprasad2008-afk · 2026-06-03T18:44:58Z

Fix crash when OpenAI-compatible providers return non-standard error payloads (e.g. OpenRouter 504s missing the choices array)
Add automatic model fallback: when a retryable error (429/504/timeout) occurs, the client transparently retries with a known-good fallback model for the same provider
Add provider presets for 9 backends: anthropic, openai, openrouter, gemini, ollama, groq, together, nvidia, mistral — each with auto- configured base URLs, default models, and fallback models
Make LLM request timeout configurable via --timeout CLI flag (default: 70000ms primary, ~45500ms compressed fallback)
Wrap each finding in a colored box border (red for HIGH, yellow for MED, gray for LOW) for improved terminal readability
Expand README with Quick Start guide, Supported Providers table with setup examples, model fallback docs, and Troubleshooting section
Add tests for undefined choices, OpenRouter error payloads, automatic model fallback, and new provider preset resolution

istg vro its good . trust me.

…tput, and enhanced README - Fix crash when OpenAI-compatible providers return non-standard error payloads (e.g. OpenRouter 504s missing the choices array) - Add automatic model fallback: when a retryable error (429/504/timeout) occurs, the client transparently retries with a known-good fallback model for the same provider - Add provider presets for 9 backends: anthropic, openai, openrouter, gemini, ollama, groq, together, nvidia, mistral — each with auto- configured base URLs, default models, and fallback models - Make LLM request timeout configurable via --timeout CLI flag (default: 70000ms primary, ~45500ms compressed fallback) - Wrap each finding in a colored box border (red for HIGH, yellow for MED, gray for LOW) for improved terminal readability - Expand README with Quick Start guide, Supported Providers table with setup examples, model fallback docs, and Troubleshooting section - Add tests for undefined choices, OpenRouter error payloads, automatic model fallback, and new provider preset resolution

ni5arga · 2026-06-03T18:46:59Z

istg vro its good . trust me.

epic comment - will review soon.

Copilot

Pull request overview

This PR strengthens the tool’s robustness and usability by improving handling of OpenAI-compatible provider quirks, adding provider presets + automatic model fallback, making timeouts configurable, and enhancing the human-readable report/README experience.

Changes:

Add provider presets for multiple OpenAI-compatible backends and extend provider parsing/validation accordingly.
Improve OpenAI-compatible client resilience (non-standard error payloads, retryable errors) and add fallback-model retry behavior with tests.
Add configurable LLM request timeouts and expand README documentation (quick start, provider setup, troubleshooting).

Reviewed changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
src/report.ts	Adds “boxed” rendering for findings with severity-colored borders/badges.
src/llm/types.ts	Expands the `Provider` union to include additional backends.
src/llm/openai.ts	Adds retryable-error detection, fallback model retry, and better handling for missing `choices`.
src/llm/openai.test.ts	Adds coverage for undefined choices and fallback behavior.
src/llm/index.ts	Introduces provider presets and improves provider normalization/error messages.
src/llm/config.test.ts	Adds tests for new provider names and preset resolution.
src/index.ts	Adds `--timeout` CLI flag and wires it into analysis options.
src/analyze.ts	Makes timeouts configurable and derives a compressed-pass timeout from the primary.
README.md	Adds Quick Start, Supported Providers, fallback docs, and Troubleshooting.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

+  const boxed = contentLines.map((line) => {
+    return color("│") + " " + line;
+  });


+    env.OPENAI_MODEL ??
+    preset?.defaultModel ??
+    "gpt-4o-mini";
+  const fallbackModels = preset?.fallbackModels ?? [];


  readonly label: string;
  readonly model: string;
+  private readonly fallbackModels: string[];
+  private activeModel: string;
  private readonly client: OpenAI;


+known-good fallback model for that provider. You'll see a warning like:
+
+```
+⚠ Model "openrouter/free" failed, falling back to "google/gemini-2.0-flash-exp:free"


Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

tejasprasad2008-afk · 2026-06-05T08:38:14Z

still dont wanna check it out @ni5arga

ni5arga · 2026-06-05T08:39:50Z

It becomes hard for me to review large PRs and this is a very medium to large sized PR (+/- 600 LoC). I'll recommend you to close this PR and make new PRs in chunks, so it'll be easier for me to review them in a proper manner.

I'll just end up running Claude Code review or Copilot on PRs of this size and merge them - that'll just degrade code optimization and I don't want to push code which I haven't manually reviewed.

Hope this makes sense.

ni5arga · 2026-06-05T08:44:43Z


> deanonymizer@0.1.0 format:check
> prettier --check "src/**/*.ts"

Checking formatting...
[error] src/llm/index.ts: SyntaxError: Argument expression expected. (89:3)
[error]   87 |   throw new Error(
[error]   88 |     `Unknown LLM provider "${value}". Supported: ${[...KNOWN_PROVIDERS].join(", ")}.`,
[error] > 89 |   if (p === "anthropic" || p === "openai" || p === "claude-code") return p;
[error]      |   ^
[error]   90 |   throw new Error(
[error]   91 |     `Unknown LLM provider "${value}". Use "anthropic", "openai", or "claude-code".`,
[error]   92 |   );
[error] src/report.ts: SyntaxError: ':' expected. (14:7)
[error]   12 |   medium: pc.bgYellow(pc.black(pc.bold("  MED "))),
[error]   13 |   low: pc.bgBlack(pc.dim(" LOW  ")),
[error] > 14 | const CONF_TAG = {
[error]      |       ^
[error]   15 |   high: pc.red(pc.bold("HIGH")),
[error]   16 |   medium: pc.yellow(pc.bold("MED")),
[error]   17 |   low: pc.dim(pc.bold("LOW")),
Error occurred when checking code style in 2 files.
Error: Process completed with exit code 2.

Also, tests are failing - you might want to check that.

tejasprasad2008-afk · 2026-06-05T08:54:15Z

Also, tests are failing - you might want to check that.

I mean it worked on my system when i ran it but def checking it rn

ni5arga · 2026-06-05T08:55:37Z

It becomes hard for me to review large PRs and this is a very medium to large sized PR (+/- 600 LoC). I'll recommend you to close this PR and make new PRs in chunks, so it'll be easier for me to review them in a proper manner.

Also, please try to take this into consideration. Thanks.

tejasprasad2008-afk · 2026-06-05T14:03:35Z

Closed in favor of breaking this down into smaller, easier-to-review PRs as requested! I'll be opening them shortly.

ni5arga requested a review from Copilot June 4, 2026 09:59

Copilot started reviewing on behalf of ni5arga June 4, 2026 09:59 View session

Copilot AI reviewed Jun 4, 2026

View reviewed changes

tejasprasad2008-afk and others added 2 commits June 4, 2026 18:11

Potential fix for pull request finding

335f87a

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

Merge branch 'main' into fix/error-handling-and-improvements

065fdbc

tejasprasad2008-afk closed this Jun 5, 2026

Navigation Menu

feat: improve error resilience, add provider presets, boxed report outtput, and enhanced README - #6

feat: improve error resilience, add provider presets, boxed report outtput, and enhanced README#6
tejasprasad2008-afk wants to merge 3 commits into
ni5arga:mainfrom
tejasprasad2008-afk:fix/error-handling-and-improvements

tejasprasad2008-afk commented Jun 3, 2026

Uh oh!

ni5arga commented Jun 3, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

tejasprasad2008-afk commented Jun 5, 2026

Uh oh!

ni5arga commented Jun 5, 2026 •

edited

Loading

Uh oh!

ni5arga commented Jun 5, 2026

Uh oh!

tejasprasad2008-afk commented Jun 5, 2026

Uh oh!

ni5arga commented Jun 5, 2026

Uh oh!

tejasprasad2008-afk commented Jun 5, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

tejasprasad2008-afk commented Jun 3, 2026

Uh oh!

ni5arga commented Jun 3, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

tejasprasad2008-afk commented Jun 5, 2026

Uh oh!

ni5arga commented Jun 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ni5arga commented Jun 5, 2026

Uh oh!

tejasprasad2008-afk commented Jun 5, 2026

Uh oh!

ni5arga commented Jun 5, 2026

Uh oh!

tejasprasad2008-afk commented Jun 5, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

ni5arga commented Jun 5, 2026 •

edited

Loading