Bohdan Triapitsyn
f4743ea060
Adds a work-status panel beside the transcript. Context fill, model and cost, todos, running subagents and the permission requests blocking them, branch and working-tree state, MCP servers, pinned messages and context sources were scattered across the header, the composer and the context panel — a blocked subagent was reported nowhere at all. The panel reads them from live channels rather than persisted history, and becomes an overlay where the chat is too narrow to seat a column. It is on by default, including for existing installs. Because it now carries these readouts, the desktop header and composer drop the ones it duplicates: todo and changed-files chips, usage and MCP tabs. VS Code and mobile keep theirs — neither hosts the panel. Fixes MCP authorization, which was broken from the panel, invalidated by a directory switch through a redirect URI that encoded the working directory, and left the desktop app in the background because browsers will not follow a custom-protocol link without a user gesture. The settings page no longer asks the user to understand the MCP spec before adding a server: one field takes the command or the link, with the kind inferred and a visible override, and client-registration fields appear only when a server actually asks for its own credentials. Also: skills load from the panel instead of only when the composer's slash autocomplete opens; the header button names the current instance rather than falling through to the word "Instance" for relay hosts. Three new optional UI settings keys, all migrated. No change to stored MCP server configuration.
75 lines
2.7 KiB
TypeScript
75 lines
2.7 KiB
TypeScript
/**
|
|
* Narrowing the discovered skill list to what the agent can actually invoke.
|
|
*
|
|
* OpenChamber scans every root it knows about. OpenCode loads a narrower set,
|
|
* governed by environment flags the browser cannot read — the server reports
|
|
* them alongside the scan.
|
|
*
|
|
* OpenCode's own skill-list endpoint cannot serve as the authority here:
|
|
* measured against 1.18.14, it returns only global and builtin skills and omits
|
|
* the project `.agents`/`.claude` skills the agent demonstrably has. Mirroring
|
|
* its discovery rules is the only way to match what the agent sees.
|
|
*
|
|
* The rules, from `opencode/src/skill/index.ts`:
|
|
*
|
|
* - `OPENCODE_DISABLE_EXTERNAL_SKILLS` drops `.claude` and `.agents` entirely;
|
|
* - `OPENCODE_DISABLE_CLAUDE_CODE` (broad) or `..._CLAUDE_CODE_SKILLS`
|
|
* (specific) drops `.claude` only — `.agents` is always scanned;
|
|
* - names are deduplicated, last scan winning, and `.agents` is scanned after
|
|
* `.claude`, so `.agents` wins a collision. This matters here because
|
|
* `.claude/skills` entries are commonly symlinks back into `.agents/skills`.
|
|
*/
|
|
|
|
type ExternalSkillFlags = {
|
|
/** `.claude` roots are not loaded. */
|
|
claudeDisabled: boolean;
|
|
/** Neither `.claude` nor `.agents` roots are loaded. */
|
|
allDisabled: boolean;
|
|
};
|
|
|
|
type SkillLike = { name: string; path: string };
|
|
|
|
const CLAUDE_ROOT = /(^|\/)\.claude\//;
|
|
const AGENTS_ROOT = /(^|\/)\.agents\//;
|
|
|
|
type SkillRoot = 'claude' | 'agents' | 'opencode';
|
|
|
|
export const resolveSkillRoot = (skillPath: string): SkillRoot => {
|
|
if (CLAUDE_ROOT.test(skillPath)) return 'claude';
|
|
if (AGENTS_ROOT.test(skillPath)) return 'agents';
|
|
return 'opencode';
|
|
};
|
|
|
|
export const filterSkillsByRuntimeFlags = <T extends SkillLike>(
|
|
skills: readonly T[],
|
|
flags: ExternalSkillFlags | null | undefined,
|
|
): T[] => {
|
|
// No flags reported means an older server or a failed read. Filtering on a
|
|
// guess would hide skills that do work, so the list passes through.
|
|
if (!flags) return [...skills];
|
|
|
|
const allowed = skills.filter((skill) => {
|
|
const root = resolveSkillRoot(skill.path);
|
|
if (root === 'opencode') return true;
|
|
if (flags.allDisabled) return false;
|
|
if (root === 'claude') return !flags.claudeDisabled;
|
|
return true;
|
|
});
|
|
|
|
// Deduplicate by name, preferring `.agents` — the same order OpenCode's
|
|
// last-write-wins scan produces.
|
|
const byName = new Map<string, T>();
|
|
for (const skill of allowed) {
|
|
const existing = byName.get(skill.name);
|
|
if (!existing) {
|
|
byName.set(skill.name, skill);
|
|
continue;
|
|
}
|
|
if (resolveSkillRoot(existing.path) === 'claude' && resolveSkillRoot(skill.path) === 'agents') {
|
|
byName.set(skill.name, skill);
|
|
}
|
|
}
|
|
|
|
return [...byName.values()];
|
|
};
|