Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add brightwave-inc/tidebreak --skill presentationsgit clone --depth 1 https://github.com/brightwave-inc/tidebreakWrote 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/brightwave-inc/tidebreak/presentations)<a href="https://agentmods.dev/skills/brightwave-inc/tidebreak/presentations"><img src="https://agentmods.dev/badge/skills/brightwave-inc/tidebreak/presentations/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/brightwave-inc/tidebreak/presentations"><img src="https://agentmods.dev/badge/skills/brightwave-inc/tidebreak/presentations.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00044 | $0.03640 |
| Opus 5.5 | $0.00018 | $0.01456 |
| Sonnet 5.5 | $0.00009 | $0.00728 |
| Haiku 4.5 | $0.00004 | $0.00364 |
Grade A, and why
presentations 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 1mo 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 — 312 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Presentations
Two paths produce a PPTX, and picking the wrong one destroys work. Choose first, then follow that path's section, and finish with the validation section — it applies to both.
Which path
- An existing
.pptxis in play — a template, a deck the user uploaded, a prior version of this deliverable, a downloaded starting point — edit it in place with the XML pipeline below. Always. The edit path is Python + the bundled helpers; it does not need Node. - Nothing exists yet and there is no template — generate a new deck. Prefer pptxgenjs when Node and npm are actually usable; fall back to python-pptx when they are not. Detect before you write the generator (see below) — do not assume either runtime from the skill name alone.
Never rebuild an existing deck to apply a change. Regenerating a deck re-authors every slide: the master, theme, fonts, colors, logos, footers, and every slide you did not think to re-emit are gone, silently, and the user sees a deck that no longer looks like theirs. A one-line text fix is a one-line XML edit, not a rewrite. Equally, never recreate from scratch what a template already provides — if the user handed you a template, its design is the requirement.
New decks: detect the runtime first
Before writing a generator, decide Node vs Python carefully. Do not spend turns
on npm install or a pptxgenjs script when node/npm are truly missing —
that is how runs burn steps and then redo the deck in Python. Also do not treat
every failed probe as "Node is gone for the turn."
- Read this turn's operating notes (the environment / capability lines
injected into the prompt). Gate the fallback on note state first:
- Node is unavailable → skip Node entirely; use the python-pptx path (install may still need package-manager network policy).
- Node is being installed → do work that does not need Node, then recheck later in the turn. Do not start the python-pptx path on the first failed probe while install is in progress. If Node still does not run after a later recheck, say so and only then take the python-pptx path (or a markdown outline with the user's knowledge) — do not silently switch generators mid-turn.
- Node is available, or the notes are silent about Node → continue to step 2. Notes can lag a managed image or a host that lost its toolchain; they do not replace a real binary check, and a probe always overrides a stale "available" claim.
- Probe the binaries once in a single
exec(no package install yet), unless step 1 already sent you to python-pptx because Node is unavailable:
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.
- 1mo ago First seen · 312 lines · 44 tokens per session scan A 33421aabadfc
presentations is a skill published in the GitHub repository brightwave-inc/tidebreak (32 stars, last pushed today), licensed Apache-2.0. It adds 44 tokens to every session and 3,640 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
ocr-and-documents
Extract text from PDFs, images, scans, Word docs (Python).
jamjet-durability-patterns
Production durability patterns for AI agents on the JVM — crash recovery, audit trails, human-in-the-loop, and replay testing using JamJet runtime.
simplicio-orient
Terminal-first execution — answer facts with the shell, never with the LLM. Use whenever a step needs a fact about the filesystem, git, processes, or system resources, or runs a build/test/lint/diff whose output would flood context. Substitutes deterministic shell/CLI calls for native LLM operations and clamps their…
simplicio-cli
Use simplicio-cli for verified code changes through mapper, precedent, six-layer prompts, tests, and a verify loop. Trigger on code edits, task-to-code, diff+test, or simplicio-py requests.
simplicio-dev-cli
Perform deterministic Simplicio code changes and validation through the Dev CLI. Use for file edits, patches, implementation, formatting, tests, diagnostics, pre-effect validation, retries, evidence files, and safe mutation workflows. The agent decides intent; Dev CLI owns the mutation and verification.
cargo-test
Skill "cargo-test" from rootwarp/claude-code-plugins-monorepo, covering /cargo-test and instructions.