Tool catalog

AI Agent Trace Failure Analysis

Given a pasted agent execution trace/log (LangChain, CrewAI, or AutoGen), returns the point and type of failure, a root-cause explanation, a suggested fix, a step-by-step summary of the run, any failure patterns detected (e.g. retry loops), and a token-usage breakdown. A trace with no failures returns a clean-run report instead. Your input is sent to an outside service for processing; remove secrets and personal data first.

AI Code Review (Bugs, Security & Style Issues)

Given a code snippet (with an optional focus area -- security, performance, readability, bugs, or all -- and an optional language hint), returns an automated review: an overall quality score, a list of issues each with a severity, category, line number, and a fix suggestion, plus a plain-language summary, quick wins, security flags, and positive aspects. Your input is sent to an outside service for processing; remove secrets and personal data first.

API Documentation Generator (from OpenAPI Spec or Plain Description)

Given either a full OpenAPI/Swagger JSON spec or a plain-language description of a single REST endpoint, returns ready-to-use markdown API documentation (description, authentication section, per-endpoint parameter tables, request/response examples, error-code tables) plus a structured summary of every endpoint documented (HTTP method, path, one-line summary, parameter count, whether it requires authentication). Your input is sent to an outside service for processing; remove secrets and personal data first.

Code Converter (Between Programming Languages)

Given a code snippet and a target language (with an optional source-language hint, auto-detected when omitted), returns the code rewritten in the target language using that language's own idiomatic patterns, along with any dependencies the converted code needs and notes explaining what changed. Your input is sent to an outside service for processing; remove secrets and personal data first.

Credential Leak Scanner for Text

Scans a block of text, such as source code, a config file or log output, line by line for credential patterns. Returns each finding's type, line number, severity, the pattern that matched and a masked excerpt (only its first and last four characters), plus a count of findings by severity (critical, high, medium), the total, a clean flag and the number of lines scanned. Pattern matching only. Your text is sent to an outside scanning service in the request address; include no live secrets.

Cron Expression Explainer (UTC)

Explains a standard five-field cron expression: returns the expression back, whether it is valid, a plain-English summary of the schedule, a breakdown of each field (minute, hour, day of month, month, day of week) with the field's value and meaning, and the next run times after the moment of the call as ISO timestamps in UTC (five by default, or three on request).

Cron Expression from Plain-English Schedule

Turns a schedule written in plain English into a standard five-field cron expression, and returns a plain-English reading of that expression and the schedule text it was built from. The expression is written by a language model, so check it before use.

CSV Delimiter Normalizer

Takes CSV text, detects its delimiter and returns the file rewritten comma-separated with quoted fields kept intact, plus the detected delimiter, the row count (header included), the column count and a list of the fixes applied. Built for files that use one delimiter throughout. Your input is sent to an outside service for processing; remove secrets and personal data first.

CVE Lookup by ID (All Sources)

Given a CVE id, returns its records from all sources the lookup covers: the National Vulnerability Database record (description, CVSS version, score, severity, vector, CWE ids, publish and modified times, reference links, page link) and, when one exists, the GitHub security advisory (GHSA id, summary, advisory text, severity, score, CWE ids, affected ecosystems, publish and update times, page link), plus the record count.

Email Address Check with Mail Server List (Format, MX, Webmail Provider)

Checks one email address: whether its format is valid, whether its domain has mail (MX) records and the mail server host names, and whether it is a disposable, role (such as contact@) or consumer-webmail address, with an overall valid flag.

Email Address Validity and Mail Host Check (Second Route)

Checks one email address and returns an overall valid flag, whether the address is well formed, its domain, whether that domain publishes mail (MX) records and the mail server host names, and whether the address is on a disposable-mailbox domain, is a role mailbox (such as info@) or is on a consumer webmail provider. Runs on a newer version of the checking service than Omnial's other mail-server email check, so each can back up the other.

Environment Variable Template Generator

Scans a piece of source code for the environment variables it reads and, with a language model, writes a ready-to-copy .env template: for each variable, a comment saying what it is and whether it is required, then a NAME=sample line. Also returns each variable as a structured entry (name, a short description, a sample value, whether the code needs it, and a line number where the code reads it) and the number of variables found. Sample values are placeholders, not real settings. Your input is sent to an outside service for processing; remove secrets and personal data first.

Fetch URL as Base64

Fetches the resource at a URL and returns its raw bytes as a base64 string, together with the content type the server reported and the size of the fetched content in bytes.

Git Commit Message Generator (from a Diff)

Given a git diff (and an optional message style: conventional, angular, or simple), returns a generated commit subject and body, the inferred change type (e.g. 'feat', 'fix') and scope, whether the change is a breaking change, and a one-line summary of the files touched. With the conventional style, full_message can repeat the type and scope prefix (feat(validate): feat(validate): ...); subject carries it once. Your input is sent to an outside service for processing; remove secrets and personal data first.

GitHub Repository Stats & Repository Search

Given one or more `owner/repo` names, returns each repository's stars, forks, open issue count, primary language, topics, license, and created/updated/pushed timestamps. Given a search query instead, returns matching repositories in the same shape, with an empty `items` array when nothing matches.

Hardcoded Secret Finder for Pasted Text

Checks a pasted block of text (code, a config file or log lines) for hardcoded secrets by pattern. For each match it returns the secret type, the line number, a severity, the pattern name and the matched text masked to its first and last four characters; it also returns the match total, per-severity counts, a clean flag and the number of lines read. Your text is sent to an outside scanning service in the request address; include no live secrets.

JavaScript Error Explainer

Explains a JavaScript error message in plain English, given the error text and a short note on where it happened: the error's type, what it means, its likely root cause, a list of suggested fixes described in words, advice on preventing it, related error messages and search terms for finding more help. Written by a language model; code samples are not returned. Your input is sent to an outside service for processing; remove secrets and personal data first.

List Supported Assistant Call Languages

One parameterless call returns a reference list of the spoken languages an AI phone-call assistant supports, each as a short code (e.g. 'en') paired with its human-readable name (e.g. 'English').

Log File Parser & Error/Warning Extractor

Given raw log text (auto-detected as JSON or syslog lines, with ISO, logfmt, and Laravel-style formats also documented by the parser though not yet seen in a captured run), returns the total line count, the detected format, a count of lines per severity level, the full list of error and warning lines pulled out on their own, a sample of parsed entries (timestamp, level, message) from the whole log, and the first/last timestamps seen. Your input is sent to an outside service for processing; remove secrets and personal data first.

Markdown to Styled HTML Converter

Given raw Markdown text and an optional theme (light, dark, or plain), returns a complete, ready-to-render HTML document with inline CSS styling, plus a bare body-only HTML fragment for embedding into an existing page. Markdown tables are not converted: each table row comes back as a plain paragraph of pipe-separated text. Your input is sent to an outside service for processing; remove secrets and personal data first.

Nginx Server Config Generator (SSL, Gzip, Rate Limiting)

Given a domain, returns a ready-to-use nginx server-block config with HTTPS/SSL, HTTP/2, WebSocket proxying, and security headers, plus a list of which features and security headers were included and setup notes for deploying it. Optional flags add gzip compression, rate limiting, a custom upstream port, and a www redirect.

Node.js Dockerfile Generator (Multi-Stage, Alpine)

Writes, with a language model, a multi-stage Dockerfile on an Alpine Node.js base image for a Node.js app built on the web framework you name. Returns the Dockerfile text, labels for its build stages, a rough image-size estimate, a list of security notes on the choices made, and recommended .dockerignore entries. The entry file, port and health check in the file are the model's assumptions; check them against your app.

NPM Package Info Lookup (Version, License, Downloads, Dependencies)

Given an npm package name, returns its latest version, license, author metadata, keyword tags, weekly download count, dependency and dev-dependency lists, npm dist-tag pointers (latest, next, beta, rc, canary, experimental, backport), the five most recently published versions, and total version count. A package name that does not resolve to a real package on the registry returns an error code and message instead of package data.

npm Package Vulnerability Lookup (by Version)

Given an npm package name and an exact version, lists the known security advisories that affect that version: for each, the advisory id, its alias ids (CVE and GHSA), a one-line summary, a severity label, the version that fixes it, the publish and last-modified times, and a few reference links. Also returns the advisory count and the highest severity found. The same or a related issue can be listed more than once under different ids; the aliases show the overlap.

Plain-English SQL Query Explainer

Given a raw SQL query, returns a plain-language explanation of what it does, a step-by-step breakdown, the tables it references, potential performance issues, and concrete optimization suggestions. Your input is sent to an outside service for processing; remove secrets and personal data first.

PyPI Package Metadata Lookup (Version, Dependencies, License, Release History)

Given a PyPI package name, returns its latest version, summary, author, license, declared dependencies, supported Python versions, and ten release versions with upload dates (an early slice of the release history that may not be the newest releases). Returns an error instead when the package name does not exist on PyPI.

Python Docstring Generator (Google, NumPy, or Sphinx Style)

Given a snippet of Python source code (a function, method, or class) and a target docstring style (Google, NumPy, or Sphinx), returns the same code with generated docstrings inserted, plus a count of how many functions and classes were documented. Your input is sent to an outside service for processing; remove secrets and personal data first.

Release Notes from Commit Messages (Changelog Sections and Markdown)

Turns a list of commit messages and a version number into release notes: a one-line summary, the change lines sorted into changelog sections (added, changed, fixed, removed, deprecated, security) and the whole notes as Markdown, plus the version echoed back and the date the notes were generated. The text is machine written, so review it before publishing.

Remote File to Ready-Made Data URI

Downloads the file at a web address and returns it ready to inline: a complete data URI (media type plus base64 content in one string), the base64 content on its own, the media type the server reported and the file size in bytes.

Security Advisories for One npm Release (npm Registry Only)

For one npm package at one exact release, lists the published security advisories that affect it. Each advisory has its id, alias ids (CVE and GHSA), a one-line summary, a severity label, the release that fixes it, its publish and last-modified times, and a few reference links. Also returns how many advisories were listed and the highest severity among them. One issue can appear more than once under different ids; the aliases show the overlap.

Stack Overflow Question Search

Searches Stack Overflow's public question and answer archive by keyword, or looks up specific questions by id, returning each question's title, full body text, tags, score, answer count, view count, whether it has an accepted answer, and the author's public display name.

Text Token Count & Cost Estimate (by LLM Model)

Given a block of text and an optional model name (e.g. gpt-4o, claude-3-haiku, gemini, mistral, llama), returns an estimated token count for that text, an estimated dollar cost at that model's per-1,000-token rate, whether the text fits inside the model's context window, and the model's maximum context size -- computed with an algorithmic estimator, not a live call to any model provider.

Web Page External Link Extractor

Fetches one web page and lists the links on it that point to a different host (a subdomain of the page's own host counts as different): each link's address as written in the page, its anchor text (null when the link has none) and its rel values (such as nofollow or noopener). Also returns the page address that was fetched and the number of links found. Links to the page's own host are not returned.

Website Security Headers Grade (A to F)

Fetches a web page and grades its HTTP security headers from A to F with a 0-100 score. Lists each security header the page sends with its value, and each missing one with a severity and a recommended fix, plus the server header and the HTTP status code.