Borrowing it
Nothing to install: this file belongs to agigante80/actual-mcp-server. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/agigante80/actual-mcp-server/main/.github/instructions/tool-files.instructions.mdgit clone --depth 1 https://github.com/agigante80/actual-mcp-serverWrote 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/instructions/agigante80/actual-mcp-server/tool-files)<a href="https://agentmods.dev/instructions/agigante80/actual-mcp-server/tool-files"><img src="https://agentmods.dev/badge/instructions/agigante80/actual-mcp-server/tool-files/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/instructions/agigante80/actual-mcp-server/tool-files"><img src="https://agentmods.dev/badge/instructions/agigante80/actual-mcp-server/tool-files.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 | $0.00882 | $0.00882 |
| Opus 5.5 | $0.00353 | $0.00353 |
| Sonnet 5.5 | $0.00176 | $0.00176 |
| Haiku 4.5 | $0.00088 | $0.00088 |
Grade A, and why
actual-mcp-server tool-files.instructions.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 20d 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.
What it actually says
Rules for MCP tool files (src/tools/*.ts)
- Tool name MUST follow
actual_{domain}_{action}snake_case convention - File name MUST match the tool name (e.g.
accounts_create.tsforactual_accounts_create) - InputSchema MUST use
z.object({...})from Zod - Use types from
CommonSchemasinsrc/lib/schemas/common.tsfor shared fields:- Dates →
CommonSchemas.date(validates YYYY-MM-DD) - Account UUIDs →
CommonSchemas.accountId - Amounts →
CommonSchemas.amountCents(integer cents, never decimal dollars)
- Dates →
- The
callfunction MUSTInputSchema.parse(args)before any other logic - NEVER wrap an
adapter.*call in a session of your own.adapter.*methods already open one, and the API mutex (withApiLockinactual-adapter.ts) is NOT reentrant, so nesting deadlocks. What you observe is a ~30s stall thenActual API operation timed out after 30000ms (ACTUAL_OP_TIMEOUT_MS), because #270 bounds each operation inside the lock and that timeout is what breaks the deadlock. Read that error as a probable nesting bug, not a slow server. - Default to calling
adapter.*methods. Do NOT reach for@actual-app/apijust to avoid a wrapper. - No tool statically imports
@actual-app/apior performs a raw write (see AGENTS.md). A batch of pure writes that must share one cycle is an adapter method (for exampleadapter.setBudgetBatch, used bybudget_updates_batch.ts) that runs its raw writes inside onequeueWriteOperation, because calling back throughadapter.*from inside a session is the nesting deadlock above. Two summary tools dynamically import the package only for theq()query builder; that is out of scope. - A read-then-write guard belongs in the ADAPTER, not the tool (#371, #376). Four tool files used to hold one inside their own
withWriteSession; all four were migrated. A guard in the tool costsretryon the reads, bypasses the observability call site, and leaves the matchingadapter.*method reachable and UNGUARDED, which is howadapter.deleteRulesat callerless while silently failing to delete a schedule-owned rule. The single-cycle property does not require the raw api:queueWriteOperationholds the lock for its whole body - When you move a guard into the adapter, fix its test's seam too. Stubbing
adapter.withWriteSessionas a pass-through stubs away the guard itself. Use_setSkipApiInitForTests(true)with the RAW api functions stubbed BEFORE the adapter import (it destructures them at module load), and assert the single-cycle property with the witness in tests/unit/helpers/write-cycle.mjs, which samples the drain counter from inside the raw read and write stubs. A bare_getWriteQueueBatchCountForTests()delta is not enough (#376): it proves only that one drain happened, so it stays green when the guard's read moves back out ofqueueWriteOperation - Error messages must be actionable: include entity type, ID, and a suggested next tool
- After creating a tool file, you MUST:
- Export it from
src/tools/index.ts - Add the name to
IMPLEMENTED_TOOLSinsrc/actualToolsManager.ts - Run
npm run buildfirst (verify-tools reads fromdist/, notsrc/) - Run
npm run verify-toolsto confirm registration
- Export it from
- To check uncovered Actual API surface before implementing:
npm run check:coverage(prints every@actual-app/apimethod against the current tool list; read-only, safe to run)
On conflict, follow the precedence in AGENTS.md; this file adds detail for its applyTo glob and must not contradict it.
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.
- 20d ago Changed · -1 lines · +9 tokens per session 11243be05c91
- 28d ago Changed · +84 tokens per session 4298274a2cf9
- 2mo ago First seen · 31 lines · 789 tokens per session scan A 5e3d67511a73
actual-mcp-server tool-files.instructions.md is an instructions file published in the GitHub repository agigante80/actual-mcp-server (74 stars, last pushed 2d ago), licensed MIT. It adds 882 tokens to every session, about $0.0035 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 instructions, from other repositories
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode copilot-instructions.md
Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).
deepseek-harness AGENTS.md
AGENTS.md instructions for deepseek-ai/deepseek-harness, covering agents.md, pre-stable apis and released session data, repository layout, commands and host sandbox failures.