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 agentmods add skills/stroland02/sync/sync-external-resourcesnpx skills add stroland02/Sync --skill sync-external-resourcesgit clone --depth 1 https://github.com/stroland02/SyncWhat 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 | $0.00065 | $0.01671 |
| Opus 5 | $0.00032 | $0.00835 |
| Sonnet 5 | $0.00013 | $0.00334 |
| Haiku 4.5 | $0.00006 | $0.00167 |
Grade A, and why
sync-external-resources 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 2d 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 — 131 lines — stays where its author put it; the contents beside it link to each section on GitHub.
External resources, attached to the milestone that needs them
Nine repositories were nominated as background reading. Loading all nine at every milestone is the failure this file exists to prevent — most are irrelevant to an API-binding engine most of the time, and a reference consulted at the wrong moment is a distraction wearing the costume of diligence.
Each entry carries a verdict, the milestone that should consult it, and the specific question it answers. Read only the row whose milestone is in hand.
Verdicts mean:
- LOAD-BEARING — a real input to a design decision. Read it before deciding.
- REFERENCE — consult when its specific question comes up. Do not read speculatively.
- SKIP — does not serve this project. Recorded with the reason so it is not re-litigated.
Verified — read from the primary source on 2026-07-27
pbakaus/impeccable — REFERENCE, M1
A design-guidance system for AI coding agents: a skill exposing 23 subcommands, roughly 60 deterministic detector rules, a browser extension for live visual iteration, and support for 13-plus agent tools. Apache 2.0.
Its subject matter is irrelevant here. Sync has no frontend. Do not read it for design advice.
Its architecture is the reason it stays on the list. Deterministic detectors carry the load; the model handles what detectors cannot express; the whole thing ships as one invocable skill that many different agent tools can consume. That is the same split Sync adopted from Open Code Review, and it is the shape the MCP graph surface is trying to be.
The question it answers: when packaging the graph surface, how does a tool present a deterministic engine to many agent clients without becoming a chat interface?
(Star count reported by the fetch looked like a misread of the page and is not recorded.)
VectifyAI/PageIndex — SKIP for specs, STEAL-THE-IDEA at M2
MIT. Builds an LLM-generated table-of-contents tree over a document, then resolves queries by
LLM tree search rather than vector similarity. PDF-oriented intake, multi-provider through
LiteLLM, default gpt-4o-2024-11-20. Claims 98.7% on FinanceBench — self-reported, no named
baseline, and the per-query model-call count is not disclosed.
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.
- 2d ago First seen · 131 lines · 65 tokens per session scan A 15bfc2d84cba
sync-external-resources is a skill published in the GitHub repository stroland02/Sync (5 stars, last pushed 6d ago), licensed Apache-2.0. It adds 65 tokens to every session and 1,671 once invoked, about $0.0003 per session on Opus 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
thermo-nuclear-code-quality-review
An extremely strict maintainability review for abstraction quality, giant files, and spaghetti-condition growth. Emend's default review skill.
migration-completeness
What a dependency migration must do before it is finished, regardless of how good the code is. Always applied; the quality skill composes on top.
PatchEvergreen Breaking Changes Analyzer
Expert skill for analyzing breaking changes, compatibility issues, and migration planning for programming libraries across multiple languages using the PatchEvergreen database.
API Contract Validator
Validate API responses against OpenAPI/Swagger specifications, JSON Schema definitions, and consumer-driven contracts to prevent breaking changes.
backend-api-versioning
Use this skill when the user says 'API versioning', 'version strategy', 'URI versioning', 'header versioning', 'content negotiation', 'accept header', 'breaking change', 'API migration', 'deprecate endpoint', 'sunset endpoint'. This skill manages API version transitions using URI, header, or content-negotiation…
Breaking Change Detector
Compares two versions of a codebase or API and flags all breaking changes with migration hints.