Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/Love-Ash/archforgenpx agentmods add skills/love-ash/archforge/archforge-pptx-lintWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/love-ash/archforge/archforge-pptx-lint)<a href="https://agentmods.dev/skills/love-ash/archforge/archforge-pptx-lint"><img src="https://agentmods.dev/badge/skills/love-ash/archforge/archforge-pptx-lint/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/love-ash/archforge/archforge-pptx-lint"><img src="https://agentmods.dev/badge/skills/love-ash/archforge/archforge-pptx-lint.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00091 | $0.04180 |
| Opus 5.5 | $0.00036 | $0.01672 |
| Sonnet 5.5 | $0.00018 | $0.00836 |
| Haiku 4.5 | $0.00009 | $0.00418 |
Grade A, and why
archforge-pptx-lint scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 18d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 212 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Archforge: PPTX quality gate for agents
Archforge is a CLI linter for built .pptx files. It inspects the file itself
(no PowerPoint needed), so it works in any headless environment. The most damaging
pptx defects are silent: fonts fall back to an OS default without any error (the
classic case is CJK text on a Latin-only font, where the coverage knowledge runs
deepest), tracking quietly wrecks character spacing, autofit shrinks text past
readability. This linter is how you catch them before a human does.
When to run
Run it EVERY time, not just when something looks wrong:
- After every build of a
.pptxyou generated or modified. - Before telling the user a deck is done. ERROR count must be 0.
- After every fix round (fixes routinely introduce new collisions).
Install and run
pip install archforge==0.11.0 # from PyPI, pinned to this release
archforge deck.pptx --profile full --fail-incomplete --json # THE agent command
# need the AI-tell rules (default is core)
archforge deck.pptx # objective defects only (core, the 0.4.0 default)
archforge deck.pptx --fail-incomplete # incomplete checks (W18) fail: use this in CI
archforge deck.pptx --fail-on-warning # WARNs fail too
archforge deck.pptx --e2-no-exemptions # E2 numeric exemptions off (full profile)
archforge deck.pptx --strict # union of the three flags above (compatibility alias)
archforge deck.pptx --ghost # per-page title list (horizontal-logic review)
archforge deck.pptx --render pages/ # + on-image contrast check (W7); needs p01.png/p02.png-named PNGs
archforge deck.pptx --skip W14,W6 # suppress specific WARNs (WARN-only; recorded in JSON)
archforge deck.pptx --lang en # report language (codes are language-independent)
archforge deck.pptx --sarif out.sarif # SARIF 2.1.0 for GitHub code scanning
archforge deck.pptx --junit out.xml # JUnit XML for Jenkins/GitLab test reports
archforge deck.pptx --write-baseline bl.json / --baseline bl.json # adopt existing decks (beta)
archforge baseline inspect bl.json # what a baseline suppresses, under which conditions
archforge fix deck.pptx -o fixed.pptx # deterministic auto-fix for E1/E2/E4 only; re-lint after
archforge deck.pptx --html report.html # annotated visual report (wireframe + finding boxes)
archforge deck.pptx --no-config # ignore config files (untrusted decks)
archforge deck.pptx --w6-sim 0.95 --w6-cluster 5 # loosen W6 for template-driven houses
archforge deck.pptx --hard-min 5 --body-min 9 --small-min 7.5 # size gate thresholds (pt)
archforge scan decks/ --profile full --json # many files/dirs/globs at once; exit 1 if any fails
archforge demo # build broken.pptx + fixed.pptx and lint both (sanity tour)
archforge skill --install # install this skill pack into ./.claude/skills
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 18d ago Changed · -1 lines b3d812007f05
- 28d ago Changed · +5 lines 2977caa450c9
- 1mo ago Changed · +2 lines f7c57e799ea0
- 1mo ago First seen · 206 lines · 91 tokens per session scan A cec6b9a97b26
archforge-pptx-lint is a skill published in the GitHub repository Love-Ash/archforge (6 stars, last pushed 20d ago), licensed MIT. It adds 91 tokens to every session and 4,180 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
themed-cn-pptx
Build, modify, and QA-verify Chinese + IP/character-themed + QR-embeddable EDITABLE PPTX decks using PptxGenJS, with SOTA AI image generation via OpenAI GPT Image 2 or Google Nano Banana Pro, locked aesthetic recipes, and a deterministic render-QA gate (CJK overflow, text overlap, editable-text, color contrast).…
ppt-master
AI-driven presentation workflow for generating editable PPTX decks and slides, reconstructing page visuals, creating reusable Brand/Style/Layout/Deck workspaces, filling native PPTX templates, and enhancing finished PPTX files. Use when the user asks to create, generate, reconstruct, regenerate, beautify, redesign…
codex-ppt
Generate visually unified image-based PPT/PPTX decks from articles, reports, papers, notes, or outlines.
impress-presentations
Create, open or edit slides: ODP default; PPTX for PowerPoint compatibility.
pptx
Read, create, and convert Microsoft PowerPoint (.pptx) presentations — extract slide text to JSON/markdown, build decks from JSON/markdown outlines, and export to PDF.
imagegen-pptx-pipeline
End-to-end stateful workflow for designing and generating editable PowerPoint decks from briefs, templates, reference decks, brand assets, data, ImageGen comps, or user-supplied slide images. Uses multi-round content/style confirmation, reviewed per-slide comps, native 4K comp preservation, conditional Real-ESRGAN…