Answer: You control tool access with the `tools`, `allowedTools`, and `disallowedTools` options, which govern availability (whether a tool appears in Claude's context) and permission (whether Claude Code approves a call). To require approval for a specific tool, set its `permission_policy` to `always_ask` so `allowedTools` stops approving that tool's calls. Status: Confirmed Entry: tool/claude-code/permissions-and-security/control-custom-tool-access Revision: 3 Digest: sha256:e8fd11968a11fb49 (changes whenever the content changes) Timeline: Created: 2026-11-13T03:02:42Z Last changed: 2026-11-20T12:16:29Z (block policy: — → always_deny) Last verified: 2026-11-20T12:16:29Z Next check: 2026-11-21T12:16:29Z Check interval: Daily Expected change: No fixed schedule Data: Field | Value | Source | As of options | tools | [1] | 2026-11-20 availability layer | Availability | [1] | 2026-11-20 permission layer | Permission | [1] | 2026-11-20 require approval policy | always_ask | [1] | 2026-11-20 block policy | always_deny | [1] | 2026-11-20 minimum version | v0.3.301 | [1] | 2026-11-20 Sources: [1] Anthropic. "Give Claude custom tools". https://code.claude.com/docs/en/agent-sdk/custom-tools.md Role: primary. Accessed: 2026-11-20T12:16:29Z Quote: "You control two layers with the `tools`, `allowedTools`, and `disallowedTools` options:" (Supports: options) Quote: "**Availability**: whether a tool appears in Claude's context." (Supports: availability layer) Quote: "**Permission**: whether Claude Code approves a call when Claude makes it." (Supports: permission layer) Quote: "Set a tool's `permission_policy` to `always_ask`, and `allowedTools` stops approving that tool's calls" (Supports: require approval policy) Quote: "Set `always_deny` instead to [block the tool]" (Supports: block policy) Quote: "The field is TypeScript-only and requires Agent SDK v0.3.301 or later." (Supports: minimum version) Published: 2026-11-20T03:17:49Z Snapshot SHA-256: b07693fa054e2dfa05152f984f16332c405b2cba4c21f6318b5de68348eeb640 (captured 2026-11-20T12:16:27Z) Verification: Method: Automated extraction from [1] with deepseek-flash (JSON output, thinking disabled). Every value was checked by code to appear verbatim in a quoted passage of [1]. Cross-check: not-available at 2026-11-20T12:16:29Z. No independent second source is configured for this entry Last successful check: 2026-11-20T12:16:29Z Last attempt: 2026-11-20T12:16:29Z Notes and limitations: Extracted automatically from a single official source (section "Control tool access"). Related: https://latestn.com/tool/claude-code/permissions-and-security/control-permissions How do I control what Claude Code is allowed to do? https://latestn.com/tool/claude-code/permissions-and-security/define-trusted-infrastructure How do I define trusted infrastructure for Claude Code auto mode? https://latestn.com/tool/claude-code/permissions-and-security/disable-permission-prompts-unattended How do I turn off permission prompts in unattended Claude Code runs? https://latestn.com/tool/claude-code/permissions-and-security/compare-sandbox-approaches Which sandbox environment options does Claude Code offer and how do they compare? https://latestn.com/tool/claude-code/permissions-and-security/configure-proxy-for-credentials How do I configure Claude Code to use a proxy for credential management? https://latestn.com/tool/claude-code/permissions-and-security/configure-security-policies What security policies should organizations configure for Claude Code? https://latestn.com/tool/ All entries in AI tools Change log: Time | Event | Change | Source 2026-11-20T12:16:29Z | changed | block policy: — → always_deny | [1] 2026-11-20T12:16:29Z | changed | block tool policy: always_deny → — | [1] 2026-11-14T12:12:05Z | changed | block tool policy: — → always_deny | [1] 2026-11-14T12:12:05Z | changed | deny policy: always_deny → — | [1] 2026-11-13T03:02:42Z | added | Entry created | — Changes for this entry as a feed: https://latestn.com/tool/claude-code/permissions-and-security/control-custom-tool-access/latest Challenge this entry: Open challenges: None How: If a value here is wrong or out of date, send what is wrong and a link to evidence. Challenges are queued and checked against the evidence; they do not change the entry by themselves. If the entry changes as a result, the change log says so. Form: https://latestn.com/dispute/tool/claude-code/permissions-and-security/control-custom-tool-access POST: https://latestn.com/dispute/tool/claude-code/permissions-and-security/control-custom-tool-access with claim, evidence_url, and optionally field and contact (form-encoded or JSON) GET: https://latestn.com/dispute/tool/claude-code/permissions-and-security/control-custom-tool-access?claim={what is wrong}&evidence_url={link} How to cite: LatestN. (2026). How do I control which custom tools Claude Code can access or require approval for? Revision 3, last verified 2026-11-20T12:16:29Z. https://latestn.com/tool/claude-code/permissions-and-security/control-custom-tool-access (accessed 2026-11-21T09:57:08Z). @misc{LatestN_tool_claude_code_permissions_and_security_control_custom_tool_access_r3, title = {How do I control which custom tools Claude Code can access or require approval for?}, author = {{LatestN}}, year = {2026}, howpublished = {\url{https://latestn.com/tool/claude-code/permissions-and-security/control-custom-tool-access}}, note = {Revision 3; digest sha256:e8fd11968a11fb49; last verified 2026-11-20T12:16:29Z; accessed 2026-11-21T09:57:08Z} } Reuse: License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/) Attribution: Copying and adapting are allowed, including commercially, if you credit LatestN and link to the entry you used. Each entry ends with a ready-made citation. URL: https://latestn.com/tool/claude-code/permissions-and-security/control-custom-tool-access