Loading ToolBench
A toolkit for augmenting GitHub Copilot with adaptive presentation generation capabilities, including PowerPoint deck creation, font management, and presentation validation
This server has severe definition quality issues. Tool names lack action verbs (toolcheck, audit_pptx, deck_spec, inspect_template, pptx_helpers are all noun-heavy or unclear). Descriptions exist but are verbose and lack clarity on WHEN to use each tool or WHAT the LLM should expect. Input schemas are visible but lack proper type constraints and enums. No output schemas are documented. Parameters have descriptions but many lack type information or constraints. The pptx_helpers tool is particularly problematic, it describes itself as 'Design-neutral python-pptx convenience wrappers' without explaining what it actually returns or how to use it. Error handling is not evident in the provided code. No security considerations are documented despite file I/O operations.
Audit a PowerPoint deck for structural and typography risks. Detects geometry overlaps, text frame issues, font problems, and other presentation quality issues
Validate the machine-readable contract for an adaptive presentation deck. Validates schema version, request parameters, canvas dimensions, font policy, language policy, speaker notes policy, and slide specifications
Inspect a PowerPoint template without exposing slide content. Returns a deterministic, content-safe profile for a PPTX/POTX file including dimensions, theme fonts, layout inventory, and fingerprints
Design-neutral python-pptx convenience wrappers for creating PowerPoint presentations. Provides utilities for deck initialization, slide creation, shapes, text formatting, and visual elements without enforcing design decisions
Probe and cache the presentation toolchain (soffice/LibreOffice, PyMuPDF, Pillow, python-pptx), then reuse its stable inventory. Detects runtime environment and required tools, caches results to ${COPILOT_CACHE_DIR:-$HOME/.copilot/cache}/adaptive-presentation/{toolchain.json,fonts.txt}
Tool names lack action verbs. 'toolcheck', 'audit_pptx', 'deck_spec', 'inspect_template', 'pptx_helpers' do not start with verbs like get_, create_, validate_, check_. LLMs cannot infer intent from these names alone.
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields. Descriptions mention what tools do but not what they return.
Input parameter types are incomplete. 'tolerance' in audit_pptx is integer but lacks min/max bounds. 'refresh' and 'require_korean_font' are boolean but lack explanation of side effects. No enums for constrained values.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 33 | 2026-07-28+ | v2 |
pptx_helpers description is vague ('Design-neutral python-pptx convenience wrappers'). Does not explain what the tool creates, what parameters control, or what output to expect. Lacks actionable guidance for LLM selection.
No error handling guidance visible. Tools that read/write files (audit_pptx, inspect_template, pptx_helpers) have no documented error responses, recovery paths, or validation messages for invalid inputs.