Session Notes: Issue #325 — Add tools subcommand

2026-03-16: Initial implementation

325-tools-subcommand based on dev

Created abm/lib/tools.py with six subcommands:

tools list — Lists all tools on the server. Supports --name, --section, and --id regex filters. Output: tab-separated ID, version, panel section, name.
tools show — Shows full JSON details for a tool (with io_details=True for input/output info).
tools inputs — Lists the inputs required by a tool in a human-readable format (name, type, required flag, label).
tools search — Searches tools by matching a query (regex) against name, ID, and description.
tools scaffold — Generates a YAML input template from a tool's input definition. Handles conditionals (shows default case active, other cases commented), sections, repeats, data/data_collection references, and all scalar types. Comments include labels, help text, allowed values.
tools run — Runs a tool with inputs from a YAML file (-f), inline key=value args, or both (CLI overrides file). Dataset references use hda:ID / hdca:ID shorthand. Supports --wait to poll until the job completes. Uses BioBlend legacy pipe-delimited key format.

_scaffold_inputs() recursively walks the tool input tree, emitting YAML with comments for each input type (conditional, section, repeat, data, boolean, select, integer, float, text, color, hidden)
_parse_value() converts string values to appropriate types (dataset refs, booleans, numbers)
_resolve_values() recursively processes a loaded YAML dict to resolve dataset reference strings
_flatten_inputs() converts nested dicts to pipe-delimited keys for BioBlend's legacy input format