Three tools with basic descriptions but critical gaps in schema completeness and parameter documentation. clerk-resume has a single required parameter (cwd) with description; clerk-index-list has no parameters; clerk-index-read has one required parameter (terms) with description. All tools return text output with no documented schema. Descriptions are present (194-250 chars) but lack actionable guidance on when to use each tool relative to others. No error recovery hints, no output structure documentation, no parameter constraints (enums, ranges). Tool names use hyphens (non-standard verb_noun pattern) and lack clarity on composition, unclear how these three tools relate or should be sequenced.

• Rename tools to verb_noun convention: clerk-resume → get_session_context, clerk-index-list → list_index_terms, clerk-index-read → read_index_terms. This makes intent explicit and parseable by LLMs.
• Document output schema for each tool. E.g. clerk-resume should return {summary_files: [{path, date, title}], transcript_files: [{path, date}]} instead of unstructured text. Use structured JSON responses.
• Add parameter constraints to clerk-index-read: specify valid term format (alphanumeric + hyphens, max 50 chars per term), max 10 terms per call, and return error 'Invalid term: X. Valid terms: [list]' when invalid.
• Enhance error messages with recovery guidance. E.g. 'Config not found at ~/.clerk/config.yaml. Run `clerk init` to create it.' or 'No index terms found. Call list_index_terms first to see available options.'
• Add dependency hints to descriptions. E.g. clerk-index-read: 'Call list_index_terms first to see available terms. Returns links to matching summaries, read them with the file-reading tool.'