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 instructions/stroland02/sync/claude-mdgit 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 | $0.02084 | $0.02084 |
| Opus 5 | $0.01042 | $0.01042 |
| Sonnet 5 | $0.00417 | $0.00417 |
| Haiku 4.5 | $0.00208 | $0.00208 |
Grade A, and why
Sync CLAUDE.md 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 — 156 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Sync
Sync watches the third-party APIs a codebase calls and opens verified pull requests when one breaks, drifts, or wastes money.
The binding is the product. The API Dependency Graph — static call sites joined against vendor
changes and runtime telemetry — is what customers pay for. Repair is a feature that proves the
binding was right, not the spine. Detectors query the graph and all emit one Finding into one
remediation pipeline. Building something that neither reads nor writes that graph? Question it.
Design: docs/superpowers/specs/2026-07-25-sync-self-maintaining-apis-design.md.
Positioning: specs/2026-07-25-sync-positioning-and-open-core.md.
The governing principle
Encode a rule where it fails, not where it is read. A rule in prose is enforced by whoever remembers it; a rule in a test, type, lint, schema constraint or hook is enforced always.
Measured in this repository: the conventions that held were machine-enforced. The ones that decayed were prose — seven of twenty-four "protected" console sentences cited deleted files and nothing noticed for two weeks. If you are about to write a paragraph asking the next agent to remember something, write the check instead.
Environment — these will break your session
| Python | 3.12. The interpreter is python. Never python3 — a Microsoft Store shim that will not run. |
| Packages | uv only. Poetry is not installed. |
| Database | Postgres 16 on port 5433. No admin rights: Docker and WSL2 are unavailable. Embedded cluster at ~/.sync-postgres — npm run no-admin starts it. Does not survive a reboot (B191). |
| Shell | Snippets are POSIX for Git Bash. PowerShell 5.1 has no && — chain with ; if ($?) { }. |
| GitHub | gh, already authenticated. |
| TypeScript | via npx. No vendored compiler. |
Never git stash when another worktree may be active. refs/stash is one ref shared across
every worktree — two agents each popped the other's stash (2026-08-16). Use git diff or a scratch
branch.
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 · 156 lines · 2,084 tokens per session scan A 84b602da2a0b
Sync CLAUDE.md is an instructions file published in the GitHub repository stroland02/Sync (5 stars, last pushed 6d ago), licensed Apache-2.0. It adds 2,084 tokens to every session, about $0.0104 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 instructions, from other repositories
onWatch CLAUDE.md
Claude Code instructions for onllm-dev/onWatch, covering onwatch, task, code map, objectives and operations.
cli AGENTS.md
Instructions for vibgrate/cli, covering agents.md, build, test, and check, prefer vg over grepping, determinism rules (non-negotiable) and public-only rule.
onWatch AGENTS.md
AGENTS.md instructions for onllm-dev/onWatch, a project described as: Track AI API quotas across Synthetic, Z.ai, Anthropic (Claude Code), Codex, GitHub Copilot & Antigravity in real time. Lightweight background daemon (<50MB RAM), SQLite storage, Material Design 3 dashboard. Zero telemetry.
mcp-ratchet AGENTS.md
Instructions for hexxla/mcp-ratchet, covering project instructions, your responsibilities as an agent, hexagonal architecture, core concepts and layer responsibilities & rules.
api-governance copilot-instructions.md
Copilot instructions for coderifts/api-governance: Call preflightchangeset before merge, deploy, publish, or tool registration when a contract artifact changed (OpenAPI/Swagger, GraphQL, gRPC/protobuf, AsyncAPI, MCP manifest, or agent tool schemas).
upgradelens AGENTS.md
AGENTS.md instructions for mattpicone/upgradelens: Remote MCP (streamable HTTP, no key for the anonymous free quota): https://upgradelens.mattpicone.workers.dev/mcp.