Pembantu
Agen pembantu persisten untuk riset, audit, kualitas, dan orkestrasi dengan kontrak frontmatter dan protokol isolasi penyebaran.
Pembantu persisten berada sebagai berkas datar di bawah direktori pembantu
kanonik dengan akhiran <name>.md. Jalur kanoniknya adalah:
src/apothem/agents/Setiap pembantu adalah peran yang terfokus, seperti riset, audit, kualitas, refactoring, atau scaffolding, dengan pembatasan tool dan system prompt-nya sendiri.
Tabel dan daftar jalur di bawah menampilkan empat dari pembantu tersebut. Inventaris yang dihasilkan di akhir halaman ini mencantumkan setiap definisi pembantu.
| Pembantu | Berkas | Pola | Kebijakan Model |
|---|---|---|---|
| Codebase Explorer | lihat jalur manifes di bawah | Riset | dipilih host |
| Convention Auditor | lihat jalur manifes di bawah | Audit | dipilih host |
| Quality Gate | lihat jalur manifes di bawah | Kualitas | dipilih host |
| Memory Auditor | lihat jalur manifes di bawah | Audit | dipilih host |
Jalur manifes kanonik:
src/apothem/agents/codebase-explorer.md
src/apothem/agents/convention-auditor.md
src/apothem/agents/quality-gate.md
src/apothem/agents/memory-auditor.mdKontrak frontmatter
---
name: <agent name>
version: <semver>
updated: <YYYY-MM-DD>
description: <one-line role>
tools: <comma-separated tool list>
disallowedTools: <comma-separated deny list>
maxTurns: <integer>
---tools bersifat opsional. Pembantu yang menghilangkannya mewarisi setiap tool
yang dimiliki sesi, termasuk tool MCP yang terhubung, dan disallowedTools
membawa batasnya. Keempat belas pembantu yang dikirimkan mencantumkan tools.
Server MCP yang terhubung
Pembantu yang mencantumkan tools hanya dapat memanggil apa yang
dicantumkannya. Claude Code menerima tool MCP ke dalam daftar itu hanya ketika
entrinya menyebut servernya, seperti mcp__<server>__*; mcp__* polos
didokumentasikan sebagai entri disallowedTools dan tidak di tempat lain. Nama
server adalah milik orang yang menghubungkannya, sehingga pembantu yang
dikirimkan tidak dapat menyebutnya.
citation-auditor dan research-scout mencari sumber ilmiah. Masing-masing
dikirimkan dengan tools: Read, Glob, Grep, WebSearch, WebFetch dan menolak tool
yang menulis berkas, menjalankan perintah, memulai pembantu lain, atau
menerbitkan: Write, Edit, MultiEdit, NotebookEdit, Bash, PowerShell,
Monitor, Agent, Artifact, dan tool daftar tugas. Tanpa pengaturan profil,
keduanya tidak menjangkau tool MCP, dan keduanya mencari sumber melalui
pencarian web dan pengambilan web.
Untuk memberi salah satunya indeks ilmiah yang terhubung, sebut servernya di
kunci agent_mcp_servers profil, yang memetakan nama agen ke nama server.
Pemasangan mesin Claude Code (apothem install) kemudian menambahkan satu entri
mcp__<server>__* per server yang disebut ke baris tools agen tersebut, dan
tidak mengubah agen lain. Pemasangan Gemini CLI melakukan hal yang sama dengan
satu entri mcp_<server>_* per server di daftar-izin tools agen, untuk nama
server yang terdiri dari huruf, angka, dan -. Di keduanya, server yang disebut
oleh disallowedTools milik agen itu sendiri tetap ditolak. Entri tersebut
mengizinkan setiap tool dari server itu, termasuk tool yang menulis, jadi sebut
hanya server yang Anda percayai untuk pencarian: instruksi pembantu membatasinya
pada panggilan search, look-up, read, verify, dan resolve, tetapi harness hanya
menegakkan nama-nama yang dicantumkan baris tools. Nama untuk agen yang tidak
ada, atau yang tidak memiliki baris tools eksplisit, dilaporkan dan dilewati.
Harness selain Claude Code dan Gemini CLI mengabaikan kunci ini. Pohon plugin
Claude Code dirakit tanpa profil dan mempertahankan daftar polos, sehingga
pengguna plugin mendefinisikan agen tingkat pengguna di bawah ~/.claude/agents/
sebagai gantinya. Sintaks profil, aturan nama server, dan solusi tersebut ada di
Memberi pembantu akses ke server MCP.
Konversi per harness
Pembantu yang pemberian tools-nya, setelah disallowedTools diterapkan, tidak
menyertakan satu pun tool penyunting berkas (Write, Edit, MultiEdit, atau
NotebookEdit) bersifat hanya-baca. Pembantu tanpa pemberian tools bersifat
hanya-baca ketika disallowedTools menolak setidaknya Write dan Edit.
Setiap harness yang mengonversi pembantu menegakkan batas itu secara native,
menyatakannya di instruksi pembantu, atau menghapus label hanya-baca:
| Harness | Batas tool | Batas giliran | Model dan tingkat upaya |
|---|---|---|---|
| Claude Code | tools / disallowedTools sebagaimana ditulis, ditambah entri mcp__<server>__* apa pun yang ditambahkan kunci agent_mcp_servers profil ke baris tools yang dicantumkan; tanpa tools, setiap tool diwarisi, termasuk MCP | maxTurns | diwarisi (tidak diatur) |
| Codex | tidak ada yang dipancarkan; instruksi pembantu menyatakannya, dan sandbox sesi induk mengatur penulisan | tidak tersedia | diwarisi (tidak diatur) |
| Gemini CLI | daftar-izin tools berisi nama tool Gemini, ditambah entri mcp_<server>_* apa pun yang ditambahkan kunci agent_mcp_servers profil ke daftar eksplisit yang dicantumkan; tanpa tools, daftar-izin yang tersirat dari penolakan ditambah mcp_* | max_turns | diwarisi (tidak diatur) |
| Qwen Code | tools / disallowedTools dalam nama tool Qwen; tanpa tools, tidak ada daftar yang dipancarkan, sehingga tool MCP diwarisi dan penolakan berlaku | tidak dipetakan | model: inherit |
| OpenCode | permission (allow / deny per tool); tool MCP tidak dibatasi | tidak dipetakan | diwarisi (tidak diatur) |
| Antigravity | tidak ada yang dipancarkan; deskripsi menghilangkan kata "read-only" | tidak tersedia | diwarisi (tidak diatur) |
| CodeBuddy | tools / disallowedTools sebagai string berkoma, terbatas pada nama tool yang didokumentasikan CodeBuddy; tanpa entri MCP | maxTurns | diwarisi (tidak diatur) |
| Trae | tools / disallowedTools sebagai string berkoma, terbatas pada nama tool yang didokumentasikan Trae; tanpa entri MCP | tidak tersedia | diwarisi (tidak diatur) |
| GitHub Copilot | daftar-izin tools berisi alias Copilot (read, search, edit, execute, web, agent, todo) dengan penolakan dilipat ke dalamnya, karena Copilot tidak memiliki kunci penolakan; alias yang mencakup tool yang ditolak dihilangkan; tanpa entri MCP | tidak tersedia | diwarisi (tidak diatur) |
| Cursor | tanpa daftar tool (Cursor tidak memilikinya); readonly: true untuk kelima pembantu hanya-baca tanpa Bash, dan pembantu hanya-baca lainnya menghilangkan label "read-only" | tidak tersedia | diwarisi (tidak diatur) |
| Kimi Code | tools / disallowedTools sebagai daftar nama tool Kimi Code (WebFetch menjadi FetchURL, TodoWrite menjadi TodoList; nama yang tidak memiliki tool di Kimi Code dihilangkan); tanpa entri MCP | tidak tersedia | diwarisi (tidak diatur) |
| Kiro | daftar-izin tools berisi kategori tool Kiro (read, write, shell, web, subagent) dengan penolakan dilipat ke dalamnya, karena daftar itu satu-satunya pembatasan yang ditulis; kategori yang mencakup tool yang ditolak dihilangkan; tanpa entri MCP | tidak tersedia | diwarisi (tidak diatur) |
| Windsurf | daftar-izin allowed-tools berisi nama tool Devin (read, edit, grep, glob, exec, write) dengan penolakan dilipat ke dalamnya, karena Devin tidak memiliki kunci penolakan; tiga pembantu yang memerlukan tool web tidak dipasang; tanpa entri MCP | tidak tersedia | model subagen bawaan Devin (tidak diatur) |
Hermes dan Open-Claw menerima berkas pembantu sebagai materi pendukung, bukan sebagai pembantu native.
Orkestrasi
Penyebaran pembantu, protokol peluncuran paralel, penegakan kontrak pengembalian, dan disiplin isolasi berada di aturan orkestrasi. Pohon dispatch per pekerja berada di aturan pola orkestrasi, §Decision Tree. Jalur kanonik:
src/apothem/rules/agent-orchestration.md
src/apothem/rules/agent-orchestration-patterns.mdInventaris Agen yang Dihasilkan
| Name | Description | Source |
|---|---|---|
README | Persistent agent definitions — the reusable sub-agent specifications the harness materializes whenever the orchestrator deploys a research, audit, or quality team. Each agent is a | src/apothem/agents/README.md |
citation-auditor | Read-only citation audit — checks that each reference in a bibliography or cited passage exists, matches its record, and that the opened source states what the text attributes to it. Use when a draft cites sources: 'check my references', 'are these DOIs real', 'does this paper say what I cite it for'. Finding new sources routes to research-scout; a free-standing claim routes to fact-checker. | src/apothem/agents/citation-auditor.md |
codebase-explorer | Read-only codebase exploration — exhaustively find every instance of a pattern, trace call/import dependencies, discover host conventions, map architecture and layering. Use when the query is 'where is X used', 'find all callers of Y', 'what convention does this repo follow for Z', 'map the module structure', or 'trace what depends on this file' — never for edits or design, and never for review verdicts (architecture-review, code-audit). Works through Grep / Glob / Read plus read-only Bash and returns evidence-cited findings (path + line + snippet) exhaustive within the declared scope; reports gaps rather than guessing. | src/apothem/agents/codebase-explorer.md |
convention-auditor | Read-only conventions auditor: verify naming (kebab-case files, snake_case Python, frontmatter keys), cross-reference resolution (dead rule/skill/agent paths, broken anchors, half-edge bindings), formatting standards (SPDX header presence, frontmatter contract), and ecosystem coherence (orphan artifacts, sibling-convention drift, registry gaps). Dispatch when a change touches multiple rules/skills/agents/commands and you need a binary PASS/FINDING verdict per instance with file+line+expected+actual evidence — e.g. 'audit naming across the harness adapters', 'check every new rule resolves its Bindings cross-refs', 'verify the skills cohort matches the kebab-case + SKILL.md convention'. Read-only: never fixes, never runs shell. A whole-tree sweep, or any fix, is the ecosystem-audit skill. Conventions are owned upstream and verified against, never re-derived; every finding cites severity + evidence. | src/apothem/agents/convention-auditor.md |
dependency-auditor | Read-only dependency-tree risk audit — flags unpinned, stale, duplicate, and known-vulnerable dependencies with manifest/lockfile evidence. Dispatch within a dependency audit, before a release cut, or after adding a dependency; one named CVE or advisory routes to the vuln-triage skill, and a signing or provenance review to the supply-chain-audit command. Detects the ecosystem via host-discovery: npm (package.json + package-lock/yarn/pnpm), pip (pyproject/requirements + lock), cargo (Cargo.toml + Cargo.lock), go (go.mod + go.sum). Runs npm audit / pip-audit / cargo audit / govulncheck and returns a risk-ranked table (known-vulnerable → unpinned → duplicate → stale) with per-row evidence; never widens a pin or escalates a scope — gaps surface as inquiry. | src/apothem/agents/dependency-auditor.md |
design-critic | Read-only UI critique. Use for a design review, a usability or accessibility check, or 'what is wrong with this screen' before a UI ships. A WCAG sweep across a whole web surface routes to the a11y-audit command; a CLI or install-flow review to ux-review. Reviews code, screenshots, or a running page and returns fixes ranked by user impact, from a scored heuristic review and separately measured evidence. Opens with a degraded notice when measurement tooling is missing; never guesses a measurement. | src/apothem/agents/design-critic.md |
fact-checker | Read-only adversarial claim verification — decompose input into atomic claims, seek ≥2 independent sources, attempt refutation, assign cited verdicts (supported / refuted / unverifiable) with quoted evidence and confidence. Use when a claim needs proof before it ships: a benchmark or statistic in docs/copy, a 'X is faster/safer than Y' assertion, a citation that names an RFC or spec, a release note, or any factual claim a reviewer would challenge. Routes external claims through WebSearch / WebFetch and repository claims through Read / Glob / Grep; defaults to refuted-or-unverifiable when evidence is insufficient, never a charitable supported. | src/apothem/agents/fact-checker.md |
mcp-builder | Scaffold a Model Context Protocol (MCP) server skeleton from a tool/resource contract — contract-first, well-typed tools, minimal surface. Use when: 'build an MCP server for <API>', 'scaffold MCP tools from this spec', 'wire FastMCP/TypeScript-SDK tool definitions', 'add a tool that exposes <resource> over MCP'. Detection: tool names + argument shapes + return types + resource URIs are stated or derivable. Selects the SDK via host-discovery (FastMCP for Python, the TypeScript SDK for Node), emits one typed tool definition per contract entry plus a list-tools smoke test, and scaffolds nothing speculative. Not for: tuning, securing, or load-testing an existing server (those surface as adjacent gaps). | src/apothem/agents/mcp-builder.md |
memory-auditor | Read-only memory-file auditor: cross-reference every claim in the harness memory tier (MEMORY.md index + topic files under <harness-root>/projects/{hash}/memory/ and <harness-root>/memory/) against actual filesystem state — file/line/rule counts (glob and count), referenced paths (do they exist?), rule scope labels (match pathFilter frontmatter?), dates (against frontmatter updated: or mtime, never the system clock), and cross-reference matrices. Dispatch when MEMORY.md or a topic file may have drifted from reality and you need a per-claim PASS/FAIL verdict with contradicting evidence — e.g. 'audit MEMORY.md after the rules cohort was renamed', 'verify the memory index counts match the current ecosystem', 'check the debugging topic file for stale references to deleted artifacts'. Existence + name match only; never re-audit an artifact's internal correctness (convention-auditor's scope). Read-only: never writes, never fixes. | src/apothem/agents/memory-auditor.md |
prompt-evaluator | Read-only rubric scoring of a prompt's output set — score each output against each named criterion (PASS/FAIL with cited evidence), aggregate per-criterion pass-rate, flag regressions against a baseline, and name recurring failure modes. Use when prompt or model outputs need a reproducible verdict before they ship: comparing a reprompted variant against the prior version, gating a system-prompt change, judging an LLM-as-judge eval set, scoring few-shot outputs against acceptance criteria, or proving a quality claim a reviewer would challenge. Scores outputs that already exist: building the dataset, scorer or runner routes to the eval-harness skill, and a full campaign to the eval command. Requires an explicit rubric — an absent or underspecified rubric blocks scoring and routes as inquiry; never invents a criterion or threshold, never scores charitably. | src/apothem/agents/prompt-evaluator.md |
quality-gate | Read-only quality-gate runner — discovers the host's lint / type-check / test / security / build commands, runs them in the correct order (build → type-check → tests+lint+security in parallel), and returns a per-gate PASS/FAIL verdict with file+line+error evidence. Reports, never fixes. Dispatch as a Quality team before a release cut, after a multi-file change, or to confirm a fix is green — e.g. 'run the full quality matrix and tell me what fails', 'gate this branch before I push', 'is the test suite green and the types clean?'. Runs the whole matrix; tests alone route to test-runner. Detects tooling via host-discovery (ruff/eslint/markdownlint, mypy/tsc/pyright, pytest/jest/cargo test/go test, bandit/npm audit/gitleaks); never assumes a stack. | src/apothem/agents/quality-gate.md |
refactor-surgeon | Scoped, behavior-preserving refactor of a named target — extract the behavioral contract, re-derive clean-room (never edit in place), name the one deficiency removed, verify regression via the host's own tests. Dispatch as a delegated team member on a single named target with a clear refactor intent; a user-driven interactive extraction is the refactor-extract skill — e.g. 'consolidate the duplicated validation in the CLI's install module', 'untangle the nested conditionals in the config renderer without changing output', 'rename the god-object methods in the adapter class to reveal intent'. Touches only the named target; adjacent gaps surface as findings, never as edits. Behavior, contracts, and side effects are identical before and after; a behavior change is a defect. | src/apothem/agents/refactor-surgeon.md |
research-scout | Read-only source discovery and ranking — decompose a research question into facets, fan out parallel web queries, rank candidates by authority, recency, and relevance, and return a deduplicated ranked source list. Use when a question needs sources before it can be answered: 'find the authoritative spec/RFC for X', 'what are the primary sources on Y', 'gather current references for a docs/copy claim', 'survey the landscape before a deep dive'. Fans external facets through any scholarly index server the operator has granted it, WebSearch, and WebFetch, and any in-repo corpus through Read / Glob / Grep. Also looks up papers and preprints by DOI, arXiv id, or topic ('find the paper behind this DOI'), and scouts existing open-source projects before anything new is built ('is there already a library for this', 'adopt or fork'). Discovery and ranking only — never fabricates a URL, never synthesizes; a written cited report routes to the multi-source-research skill, claim verification to fact-checker, and combining given documents to source-synthesis. | src/apothem/agents/research-scout.md |
security-scanner | Read-only secret, SAST-pattern, and config-risk sweep — a coarse first-pass that surfaces candidates fast and routes deep scanning to the host's CI scanners. Dispatch inside a security audit, after touching auth/crypto/IO surfaces, or when a secret may have been committed; a full repository security audit routes to the security-audit command. Greps for credential literals (key/token/password/private-key/certificate, committed .env values), injection surfaces (shell=True on untrusted input, string-formatted SQL, command interpolation), unsafe-eval (eval/exec/Invoke-Expression), unsafe deserialization (pickle.loads, yaml.load without SafeLoader), and config risk (over-broad CI permissions:, unpinned Actions uses:, network egress). Routes known-vulnerable dependencies to the dependency-auditor agent, CVE/advisory triage to the vuln-triage skill, and leaked-credential rotation to the secret-rotation skill. | src/apothem/agents/security-scanner.md |
test-runner | Read-only test-suite runner — discovers the host's test command, runs it, and triages every failure by root cause with test+file+line+assertion evidence. Reports, never fixes. Dispatch as a Research/Quality team member after a code change, before a release cut, or to confirm a fix is green — e.g. 'run the tests and tell me what's failing and why', 'is the suite green on this branch?', 'triage the failures in the auth module'. Tests only: lint, type and security gates route to quality-gate. Detects the runner via host-discovery (pytest/jest/cargo test/go test/Makefile target from the manifest + CI config); never assumes pytest. | src/apothem/agents/test-runner.md |
Perintah
Definisi slash-command yang mengimplementasikan pipeline plan-suite beserta audit untuk tinjauan kode, arsitektur, keamanan, dependensi, dan dokumentasi.
Hook
Penangan peristiwa tool yang terpicu pada SessionStart, PreToolUse, PostToolUse, dan Stop, memuat konteks pesan yang menegakkan isolasi git dan disiplin rencana, dengan kontrak runtime yang teruji dan status per pengguna.