Where AI Builds Agent Skills

A skill that generates a structured CHANGELOG.md from a list of conventional commit messages. Features: - Accept list of commit objects: {sha, message, date, author} - Group commits by type: Features, Bug Fixes, Performance, Documentation, etc. - Filter out chore/style/test commits from user-facing changelog - Format as Keep a Changelog (keepachangelog.com) style Markdown - Support version tagging: group commits between version tags - Include breaking changes section prominently Input: list of commit dicts + version string Output: Markdown changelog string

A skill that parses and filters structured log lines from common log formats. Features: - Parse log lines from formats: JSON logs, logfmt, Apache/Nginx combined log, Python logging - Filter by log level: DEBUG, INFO, WARNING, ERROR, CRITICAL - Filter by time range (start/end datetime) - Extract fields: timestamp, level, message, logger name, extra fields - Aggregate: count by level, count by logger, error rate over time - Output filtered logs as list of parsed dicts Input: log text (multiline string) + filter options Output: list of parsed log entry dicts

A skill that resolves and validates package version constraints. Features: - Parse version constraint strings: ==1.2.3, >=1.0,<2.0, ~=1.4, ^1.2.3 (npm-style) - Check if a version satisfies a constraint - Find the highest version from a list that satisfies a constraint - Detect conflicting constraints between two requirements - Support both PEP 440 (Python) and semver (npm/cargo) constraint syntax Input: version string + constraint string Output: bool (satisfies) or resolved version string

A skill that computes cyclomatic complexity and other code metrics for Python functions. Features: - Parse Python function source code - Compute cyclomatic complexity (McCabe metric) - Count lines of code (total, blank, comment, code) - Detect deeply nested blocks (configurable threshold) - Identify long functions (configurable line threshold) - Return per-function metrics and overall file summary Input: Python source code string Output: list of {function_name, complexity, loc, max_nesting_depth, warnings: list[str]}

A skill that converts datetimes between timezones and answers timezone-related queries. Features: - Convert datetime from one timezone to another using IANA timezone names - List all available timezones, optionally filtered by region - Get current time in any timezone - Calculate UTC offset for a timezone at a given datetime (handles DST) - Find overlapping business hours between two timezones Input: datetime string + source timezone + target timezone Output: converted datetime string with offset

A skill that computes edit distance and similarity between strings. Features: - Compute Levenshtein (edit) distance between two strings - Compute normalized similarity score (0.0 to 1.0) - Support Damerau-Levenshtein (includes transpositions) - Find closest match from a list of candidates - Batch compare one string against many - Configurable costs for insert, delete, substitute operations Input: two strings (or one string + list of candidates) Output: distance integer and similarity float