langchain-ai/Deep Agents Code

Run dcode in GitHub workflows

View on GitHub

Trust Signals

Scorecard Score
not yet scored
Maintenance Recency
Activelast commit Jul 10, 2026
License
MIT

Pinned Snippet

workflow.ymlSHA-pinned
uses: langchain-ai/deepagents@7e70065200007896336f38fe905803e6763e8f85 # deepagents==0.6.12

tags can be moved; commit SHAs can't. why a SHA?

namedescriptionrequireddefault
promptThe prompt/instruction to send to the agentyes
modelModel to use, as provider:model (e.g. openai:gpt-5.5, anthropic:claude-opus-4-6) or a bare name (claude-*, gpt-*, gemini-*) with the provider auto-detected.no
model_paramsJSON model kwargs passed to --model-paramsno""
max_retriesMaximum transient model retries passed to --max-retriesno""
profile_overrideJSON model profile overrides passed to --profile-overrideno""
anthropic_api_keyAnthropic API keyno
openai_api_keyOpenAI API keyno
google_api_keyGoogle API keyno
github_tokenGitHub token for API accessno${{ github.token }}
working_directoryWorking directory for the agentno.
cli_versiondeepagents-code version (empty = latest)no""
skills_repoGitHub repo of skills to clone (e.g. owner/repo, owner/repo@ref, or full URL)no""
enable_memoryPersist agent memory across workflow runs using actions/cache. When enabled, memory is keyed by agent_name + memory_scope so the agent can recall prior context.notrue
memory_scopeCache scope: pr (shared per PR), branch (shared per branch), repo (shared across repo)norepo
agent_nameAgent identity name — controls memory namespace (default: agent)noagent
shell_allow_listComma-separated shell allow list passed to --shell-allow-list (default: recommended,git,gh)norecommended,git,gh
skillSkill name to invoke before running the promptno""
startup_cmdShell command to run at startup before the promptno""
sandboxSandbox provider passed to --sandbox (empty = local execution)no""
sandbox_idExisting sandbox ID passed to --sandbox-idno""
sandbox_snapshot_nameSnapshot or blueprint name passed to --sandbox-snapshot-nameno""
sandbox_setupSetup script path passed to --sandbox-setupno""
mcp_configMCP config path passed to --mcp-configno""
no_mcpPass --no-mcp to disable MCP loadingnofalse
trust_project_mcpPass --trust-project-mcp to skip project MCP approvalnofalse
interpreterSet to true for --interpreter, false for --no-interpreter, empty for dcode defaultno""
interpreter_toolsPTC allowlist passed to --interpreter-toolsno""
max_turnsMaximum agentic turns passed to --max-turnsno""
task_timeoutdcode task timeout in seconds passed to --timeoutno""
rubricAcceptance criteria passed to --rubric (literal text or @path)no""
rubric_modelModel passed to --rubric-modelno""
rubric_max_iterationsMaximum rubric grader iterations passed to --rubric-max-iterationsno""
quietPass --quiet for clean stdoutnofalse
no_streamPass --no-stream to buffer the full responsenofalse
jsonPass --json for machine-readable dcode outputnofalse
stdinPass --stdin and pipe prompt as standard input instead of using --non-interactivenofalse
timeoutMaximum action runtime in minutes (default: 30)no30
namedescription
responseFull text response from the agent
exit_codeExit code from the agent
cache_hitWhether agent memory was restored from cache (empty if enable_memory is false)