opencode already reads skills from ~/.claude/skills and ~/.agents/skills, so
those needed nothing. The rest did.
The manifest now links agents/AGENTS.md to ~/.config/opencode/AGENTS.md as
well, beside the Claude Code and Codex names.
opencode keeps its MCP servers in its own JSON config, with no file to point
a symlink at, so link-mcp now writes them there too. It builds one entry per
row of user/agents/mcp/servers with the same tokens it gives Claude Code,
and adds ~/.claude/rules/*.md under instructions, since opencode does not
read that directory by itself. Only those two keys change. The file is
written private and swapped in whole because it now holds tokens, and a
config with JSONC comments is left alone with a warning rather than
rewritten without them.
link-mcp gains a stdio transport: the servers row names a launcher under user/agents/mcp/, run from the checkout so every machine runs the same script, and the token reaches it as MCP_TOKEN. sentry-gbrown is the first, pointing Sentry's MCP server at sentry.gbrown.org with a read-only token from the ignored env file.
A server already registered exactly as its row would register it is now left alone. Re-adding it every run threw away the sign-in of the OAuth servers, bill-tracker and atlassian, on every update.
Penpot's MCP server reads the user key only from the userToken query
parameter, never a header. link-mcp now substitutes a {TOKEN} placeholder
in a server's URL instead of sending the Authorization header, and the
servers list gains a penpot row that uses it.
An MCP server is a URL plus a bearer token, and the token is why this is a
stage rather than a manifest line. Panama is public, so the tokens cannot
live in it, and neither runtime keeps its server list in a file worth
symlinking: Codex writes them into config.toml beside dozens of unrelated
settings, and Claude Code into ~/.claude.json. link-mcp registers them
through the runtime's own CLI instead.
user/agents/mcp/servers is tracked and names which variable carries each
token. user/agents/mcp/env holds the tokens and is ignored. A new machine
gets the servers by dropping its own env file beside the tracked one.
Only Claude Code is handled, and only where the interview enabled personal
content. Rewriting a section of somebody's live Codex TOML is a worse
failure mode than leaving two lines to paste once.