langchain-ai/deepagents · error
Error: Source agent '{source_agent}' not found or has no AGE
Error message
Error: Source agent '{source_agent}' not found or has no AGENTS.md
Available agents: dcode agents list What it means
reset_agent can re-seed an agent from another agent's AGENTS.md. When the named source agent's directory (or its AGENTS.md) does not exist, it prints this Rich-formatted error including a hint to run 'dcode agents list' and raises SystemExit(1) to terminate the CLI with a failure status.
Source
Thrown at libs/code/deepagents_code/agent.py:1370
output_format: Output format — `'text'` (Rich) or `'json'`.
Raises:
SystemExit: If the source agent is not found.
"""
agents_dir = user_deepagents_dir()
agent_dir = agents_dir / agent_name
if source_agent:
source_dir = agents_dir / source_agent
source_md = source_dir / "AGENTS.md"
if not source_md.exists():
console.print(
f"[bold red]Error:[/bold red] Source agent '{source_agent}' not found "
"or has no AGENTS.md\n"
" Available agents: dcode agents list"
)
raise SystemExit(1)
source_content = source_md.read_text()
action_desc = f"contents of agent '{source_agent}'"
else:
source_content = get_default_coding_instructions()
action_desc = "default"
if dry_run:
if output_format == "json":
from deepagents_code.output import write_json
write_json(
"reset",
{
"agent": agent_name,
"reset_to": source_agent or "default",
"path": str(agent_dir),
"dry_run": True,View on GitHub (pinned to a1af029e6e)
Solutions
- Run the listed command to see valid names: dcode agents list, then retry with an exact name.
- If the agent should exist, recreate it (dcode agents create <name>) and ensure its AGENTS.md is present before resetting from it.
- If resetting to defaults was intended, omit the source-agent argument so reset_agent uses get_default_coding_instructions().
Example fix
// before dcode agents reset --from clsud # typo // after dcode agents list dcode agents reset --from cloud
Defensive patterns
Strategy: validation
Validate before calling
import subprocess
listing = subprocess.run(
["dcode", "agents", "list"], capture_output=True, text=True, check=True
).stdout
if source_agent not in listing:
raise SystemExit(f"Unknown source agent {source_agent!r}; run 'dcode agents list'") Type guard
null
Try / catch
try:
reset_agent(source_agent=name)
except SystemExit as exc:
if exc.code != 0 and "not found or has no AGENTS.md" in captured_output:
print(f"'{name}' is not a valid agent; run 'dcode agents list'")
else:
raise Prevention
- Run 'dcode agents list' before resetting to confirm the exact source-agent name.
- Check for the directory and its AGENTS.md before invoking: from pathlib import Path; assert Path('.dcode/agents', name, 'AGENTS.md').exists() (adjust to your agents root).
- Remember the check is name-exact: watch for typos and case mismatches.
- If you want defaults, call reset_agent without a source agent instead of naming a nonexistent one.
When it happens
Trigger: dcode agents reset --from <name> (reset_agent with source_agent=<name>) where <name> is not an existing agent directory or its AGENTS.md is missing — typo'd name, agent deleted, or agent created in a different checkout/namespace.
Common situations: Typo in the agent name; resetting from an agent that was removed with a partial delete (directory gone or AGENTS.md deleted); running in a different project where the source agent was never created; case-mismatched name.
Understand the failure class
Background: 'Could not be found', 'does not exist', 'not found in database': the resource-not-found family when an ID, slug, key, or URI lookup comes back empty — this error's family across 20 libraries.
Related errors
- dcode: {message}
- Could not resolve {label} path {raw_path!r}: {exc}. Ensure t
- enable_interpreter=True is not supported with a remote sandb
- --model-params requires a JSON object value
- Unclosed {quote} in --model-params value
AI-assisted analysis of langchain-ai/deepagents@a1af029e6e (2026-08-29).
Data as JSON: /api/errors/f16c8a06fc075fe7.
Report an issue: GitHub.