feat: add claude-agent-sdk single-agent and multi-agent patterns with READMEs

[iprivit]

Summary

• Two new claude-agent-sdk agent patterns alongside strands and langgraph:
  • claude-agent-sdk-single-agent: Standalone agent with Code Interpreter via in-process MCP server, Gateway integration, session management via claude_session_id, and secure JWT-based user identity extraction
  • claude-agent-sdk-multi-agent: Extends the single-agent pattern with subagent delegation — includes a code-analyst subagent for Task-based delegation via the Task tool
• Dedicated frontend SSE parser shared by both claude-agent-sdk patterns
• CDK updates: CLAUDE_CODE_USE_BEDROCK env var for both patterns and validation preventing zip deployment (requires Node.js + claude-code CLI)
• Security hardening: STACK_NAME format validation, payload field validation, graceful gateway failure handling
• Server-side session map for claude-agent-sdk conversation memory — maps runtimeSessionId to claude_session_id on the backend, no frontend round-trip needed
• Built-in tool lockdown: Disables claude-code built-in tools (Bash, Write, Read, Edit, NotebookEdit, WebFetch, Glob, Grep, EnterWorktree, Skill, TodoWrite, CronCreate, CronDelete, CronList) via disallowed_tools so the agent operates exclusively through Code Interpreter and Gateway MCP tools. Documentation explains how to re-enable individual tools.
• Pinned dependency versions: claude-agent-sdk==0.1.48 and @anthropic-ai/claude-code@2.1.73 for reproducible builds
• READMEs for all patterns (strands, langgraph, claude-agent-sdk-single-agent, claude-agent-sdk-multi-agent) documenting features, architecture, streaming events, tool configuration, memory, security, and deployment

Changed files

Single-agent pattern (patterns/claude-agent-sdk-single-agent/):

• agent.py — main entrypoint with ClaudeSDKClient, RequestContext security, Code Interpreter MCP, Gateway, server-side session map, disallowed_tools configuration
• code_int_mcp/ — in-process MCP server wrapping AgentCore Code Interpreter API (server, client, models)
• Dockerfile — Python 3.11 + Node.js + @anthropic-ai/claude-code@2.1.73
• requirements.txt — claude-agent-sdk==0.1.48, bedrock-agentcore, pydantic, PyJWT, etc.
• README.md — full pattern documentation including built-in tool configuration

Multi-agent pattern (patterns/claude-agent-sdk-multi-agent/):

• Same as single-agent, plus:
• agents/subagents.py — code-analyst subagent definition via AgentDefinition
• agent.py — adds Task tool and subagent orchestration to the single-agent base
• README.md — full pattern documentation including subagent usage and built-in tool configuration

Frontend:

• parsers/claude-agent-sdk.ts — dedicated parser for data, current_tool_use, claude_session_id events
• types.ts — added "claude-agent-sdk-single-agent" and "claude-agent-sdk-multi-agent" to AgentPattern union
• client.ts — registered both patterns with the shared parser

Shared utilities:

• patterns/utils/auth.py — added get_gateway_access_token() with @requires_access_token decorator

Infrastructure:

• backend-stack.ts — conditional CLAUDE_CODE_USE_BEDROCK=1 env var + zip deployment validation for both patterns

Documentation:

• patterns/strands-single-agent/README.md
• patterns/langgraph-single-agent/README.md

[kaleko]

I understand the differences, but my main question is why there are differences? For example why didn't we make the claude-sdk example use AgentCore Memory, single agent, and why we need an in-process MCP server for this? Would be useful to explain why the differences in the docs/code

[kaleko]

This is an interesting dependency... definitely needed?

[kaleko]

Do we want to make this claude-agent-sdk-single-agent instead of having multiple agents, so it is more apples to apples with the existing patterns? Then we can include a readme or commented code showing how to extend it to multiple agents?

[github-actions]

Latest scan for commit: e80f5ae | Updated: 2026-03-11 19:28:22 UTC

Security Scan Results

Scan Metadata

• Project: ASH
• Scan executed: 2026-03-11T19:28:06+00:00
• ASH version: 3.2.2

Summary

Scanner Results

The table below shows findings by scanner, with status based on severity thresholds and dependencies:

Column Explanations:

Severity Levels (S/C/H/M/L/I):

• Suppressed (S): Security findings that have been explicitly suppressed/ignored and don't affect the scanner's pass/fail status
• Critical (C): The most severe security vulnerabilities requiring immediate remediation (e.g., SQL injection, remote code execution)
• High (H): Serious security vulnerabilities that should be addressed promptly (e.g., authentication bypasses, privilege escalation)
• Medium (M): Moderate security risks that should be addressed in normal development cycles (e.g., weak encryption, input validation issues)
• Low (L): Minor security concerns with limited impact (e.g., information disclosure, weak recommendations)
• Info (I): Informational findings for awareness with minimal security risk (e.g., code quality suggestions, best practice recommendations)

Other Columns:

• Time: Duration taken by each scanner to complete its analysis
• Action: Total number of actionable findings at or above the configured severity threshold that require attention

Scanner Results:

• PASSED: Scanner found no security issues at or above the configured severity threshold - code is clean for this scanner
• FAILED: Scanner found security vulnerabilities at or above the threshold that require attention and remediation
• MISSING: Scanner could not run because required dependencies/tools are not installed or available
• SKIPPED: Scanner was intentionally disabled or excluded from this scan
• ERROR: Scanner encountered an execution error and could not complete successfully

Severity Thresholds (Thresh Column):

• CRITICAL: Only Critical severity findings cause scanner to fail
• HIGH: High and Critical severity findings cause scanner to fail
• MEDIUM (MED): Medium, High, and Critical severity findings cause scanner to fail
• LOW: Low, Medium, High, and Critical severity findings cause scanner to fail
• ALL: Any finding of any severity level causes scanner to fail

Threshold Source: Values in parentheses indicate where the threshold is configured:

• (g) = global: Set in the global_settings section of ASH configuration
• (c) = config: Set in the individual scanner configuration section
• (s) = scanner: Default threshold built into the scanner itself

Statistics calculation:

• All statistics are calculated from the final aggregated SARIF report
• Suppressed findings are counted separately and do not contribute to actionable findings
• Scanner status is determined by comparing actionable findings to the threshold

Scanner	S	C	H	M	L	I	Time	Action	Result	Thresh
bandit	0	0	0	0	0	0	865ms	0	PASSED	MED (g)
cdk-nag	0	0	0	0	0	0	30.9s	0	PASSED	MED (g)
cfn-nag	0	0	0	0	0	0	9ms	0	PASSED	MED (g)
checkov	0	0	0	0	0	0	4.6s	0	PASSED	MED (g)
detect-secrets	0	0	0	0	0	0	828ms	0	PASSED	MED (g)
grype	0	0	0	0	0	0	52.4s	0	PASSED	MED (g)
npm-audit	0	0	0	0	0	0	165ms	0	PASSED	MED (g)
opengrep	0	9	0	0	0	0	29.0s	9	FAILED	MED (g)
semgrep	0	9	0	0	0	0	11.3s	9	FAILED	MED (g)
syft	0	0	0	0	0	0	2.0s	0	PASSED	MED (g)

Detailed Findings

Show 18 actionable findings

Finding 1: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: semgrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:126

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const patternDir = path.join(repoRoot, "patterns", pattern)

---

Finding 2: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: semgrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:154

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const content = fs.readFileSync(path.join(patternDir, file))

---

Finding 3: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: semgrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:154

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const content = fs.readFileSync(path.join(patternDir, file))

---

Finding 4: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: semgrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:168

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const requirementsPath = path.join(patternDir, "requirements.txt")

---

Finding 5: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: semgrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:998

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const fullPath = path.join(dirPath, entry.name)

---

Finding 6: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: semgrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:998

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const fullPath = path.join(dirPath, entry.name)

---

Finding 7: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: semgrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:999

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const relativePath = path.join(prefix, entry.name)

---

Finding 8: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: semgrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:999

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const relativePath = path.join(prefix, entry.name)

---

Finding 9: python.jwt.security.unverified-jwt-decode.unverified-jwt-decode

• Severity: HIGH
• Scanner: semgrep
• Rule ID: python.jwt.security.unverified-jwt-decode.unverified-jwt-decode
• Location: patterns/utils/auth.py:65

Description:
Detected JWT token decoded with 'verify=False'. This bypasses any integrity checks for the token which means the token could be tampered with by malicious actors. Ensure that the JWT token is verified.

Code Snippet:

options={"verify_signature": False},

---

Finding 10: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: opengrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:126

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const patternDir = path.join(repoRoot, "patterns", pattern)

---

Finding 11: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: opengrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:154

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const content = fs.readFileSync(path.join(patternDir, file))

---

Finding 12: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: opengrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:154

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const content = fs.readFileSync(path.join(patternDir, file))

---

Finding 13: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: opengrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:168

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const requirementsPath = path.join(patternDir, "requirements.txt")

---

Finding 14: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: opengrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:998

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const fullPath = path.join(dirPath, entry.name)

---

Finding 15: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: opengrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:998

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const fullPath = path.join(dirPath, entry.name)

---

Finding 16: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: opengrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:999

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const relativePath = path.join(prefix, entry.name)

---

Finding 17: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal

• Severity: HIGH
• Scanner: opengrep
• Rule ID: javascript.lang.security.audit.path-traversal.path-join-resolve-traversal.path-join-resolve-traversal
• Location: infra-cdk/lib/backend-stack.ts:999

Description:
Detected possible user input going into a path.join or path.resolve function. This could possibly lead to a path traversal vulnerability, where the attacker can access arbitrary files stored in the file system. Instead, be sure to sanitize or validate user input first.

Code Snippet:

const relativePath = path.join(prefix, entry.name)

---

Finding 18: python.jwt.security.unverified-jwt-decode.unverified-jwt-decode

• Severity: HIGH
• Scanner: opengrep
• Rule ID: python.jwt.security.unverified-jwt-decode.unverified-jwt-decode
• Location: patterns/utils/auth.py:65

Description:
Detected JWT token decoded with 'verify=False'. This bypasses any integrity checks for the token which means the token could be tampered with by malicious actors. Ensure that the JWT token is verified.

Code Snippet:

options={"verify_signature": False},

---

Report generated by Automated Security Helper (ASH) at 2026-03-11T19:28:00+00:00

[kaleko]

I think all of the changes in this file can be reverted in lieu of latest force-push to main to fix frontend bugs.

[kaleko]

I think all of the changes in this file can be reverted in lieu of latest force-push to main to fix frontend bugs.

[kaleko]

I think all of the changes in this file can be reverted in lieu of latest force-push to main to fix frontend bugs. I believe the exact same fix was already pushed.