Add GitHub Copilot CLI support - #534
Merged
rohitg00 merged 14 commits intoMay 28, 2026
Merged
Conversation
- plugin/.plugin/plugin.json: Copilot manifest with name/version/skills/mcpServers/hooks refs - plugin/.mcp.copilot.json: MCP server config with type:local, npx, env passthrough, tools:[*] - plugin/hooks/hooks.copilot.json: Copilot hooks (version:1) with 11 supported events and PreToolUse matcher - test/copilot-plugin.test.ts: 11 tests covering manifest, MCP config, and hooks validation Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Adds Copilot CLI support through a root plugin manifest, Copilot-specific MCP and hook configuration, and a connect adapter for MCP-only setup. Includes Windows-safe Copilot MCP command generation, COPILOT_HOME handling, Copilot hook payload normalization, generated hook scripts, and targeted tests for plugin shape, hook execution, and connect behavior. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
|
Someone is attempting to deploy a commit to the rohitg00's projects Team on Vercel. A member of the Team first needs to authorize it. |
Contributor
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
📝 WalkthroughWalkthroughAdds GitHub Copilot CLI support: new Copilot plugin manifests (MCP + hooks), normalizes hook inputs (snake_case/camelCase) across scripts and TS hooks, implements a ChangesCopilot CLI Integration
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 6
🧹 Nitpick comments (2)
plugin/scripts/subagent-start.mjs (1)
26-28: ⚡ Quick winMixed fallback operators within the same extraction block.
Line 26 uses
||forsessionId, while lines 27-28 use??for agent fields. This inconsistency means:
- Empty string
sessionIdwould fall back to"unknown"- Empty string
agentIdoragentTypewould be preservedFor predictable behavior, use the same operator consistently. Since
??better represents "use if provided" semantics for optional fields, consider applying it to all three lines.♻️ Suggested fix
-const sessionId = data.session_id || data.sessionId || "unknown"; -const agentId = data.agent_id ?? data.agentName; -const agentType = data.agent_type ?? data.agentDisplayName ?? data.agentName; +const sessionId = data.session_id ?? data.sessionId ?? "unknown"; +const agentId = data.agent_id ?? data.agentName; +const agentType = data.agent_type ?? data.agentDisplayName ?? data.agentName;🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@plugin/scripts/subagent-start.mjs` around lines 26 - 28, The extraction uses inconsistent fallback operators: sessionId uses || while agentId and agentType use ??, causing empty strings to behave differently; update the sessionId assignment (symbol: sessionId) to use the nullish coalescing operator like agentId and agentType (symbols: agentId, agentType) — e.g., replace the || fallback with ?? so sessionId = data.session_id ?? data.sessionId ?? "unknown" to ensure consistent "use if provided" semantics across all three fields.plugin/scripts/prompt-submit.mjs (1)
25-25: ⚡ Quick winInconsistent fallback operators:
||vs??.Line 25 uses
||forsessionIdfallback, while line 36 uses??forpromptfallback. This creates different behavior:
||treats empty strings as falsy and falls back??only falls back on null/undefined, preserving empty stringsFor consistent behavior across all normalized fields, use the same operator. Consider
??throughout, as it better expresses "use this field if provided, otherwise try the alternative."♻️ Suggested fix for consistency
-const sessionId = data.session_id || data.sessionId || "unknown"; +const sessionId = data.session_id ?? data.sessionId ?? "unknown";Also applies to: 36-36
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@plugin/scripts/prompt-submit.mjs` at line 25, The fallback for sessionId uses the || operator which treats empty strings as falsy while prompt uses ??; change the sessionId normalization to use nullish coalescing (replace the || chain with ??) so it matches the prompt behavior (i.e., const sessionId = data.session_id ?? data.sessionId ?? "unknown") and scan other normalized fields in this module to ensure all use ?? for consistent null/undefined-only fallbacks.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@plugin/hooks/hooks.copilot.json`:
- Around line 16-21: preToolUse matcher is missing "write" and "read" so
pre-tool-use.mjs won't run for those actions; update the "preToolUse" entry's
matcher string (the one currently set to "edit|create|view|glob|grep") to
include "write" and "read" (e.g., "edit|create|view|glob|grep|write|read") so it
matches the allowlist used by pre-tool-use.mjs and ensures context injection for
the write/read tool calls.
In `@src/hooks/notification.ts`:
- Around line 32-35: The code incorrectly casts session values with "as string"
and mixes || with ??; update the sessionId logic to use the same nullish
coalescing pattern as notificationType and explicitly coerce the result to a
string (e.g., use data.session_id ?? data.sessionId ?? "unknown" and then wrap
with String(...) or template interpolation) so sessionId is always a string;
reference the notificationType variable and the sessionId assignment and the
data.session_id/data.sessionId fields when making the change.
In `@src/hooks/pre-tool-use.ts`:
- Around line 53-63: The code assumes data.tool_name/data.toolName is a string
and calls toLowerCase on it and similarly assumes data.tool_input/toolArgs is an
object; add runtime narrowing: for toolName, check typeof data.tool_name ===
"string" or typeof data.toolName === "string" before assigning and calling
toLowerCase (otherwise bail/return), and for toolInput ensure it's an object
(e.g., typeof === "object" && data.tool_input !== null) or default to an empty
Record<string, unknown>; update usages of toolName, normalizedToolName, and
toolInput to rely on these checks so toLowerCase and object operations are safe.
In `@src/hooks/subagent-start.ts`:
- Around line 41-43: The fallback logic is inconsistent: sessionId uses ||
(which treats empty string as falsy) while agentId and agentType use ?? (which
does not treat empty strings as nullish), so empty values like "" won't fall
back to agentName/agentDisplayName; update the assignments for agentId and
agentType in src/hooks/subagent-start.ts to use the same logical OR fallback as
sessionId (replace ?? with ||) so empty strings correctly fall back to the
alternative fields (refer to the variables agentId and agentType).
In `@src/hooks/subagent-stop.ts`:
- Around line 33-35: The fallbacks for sessionId, agentId, and agentType are
inconsistent: sessionId uses || (treats empty string as missing) while agentId
and agentType use ?? (does not treat empty string as missing). Update agentId
and agentType to use the same falsy fallback as sessionId by replacing the
nullish coalescing (??) with logical OR (||) so empty strings fall back to the
alternate fields/values; refer to the variables sessionId, agentId, and
agentType in src/hooks/subagent-stop.ts when making the change.
In `@test/copilot-plugin.test.ts`:
- Around line 171-182: The test allows empty or null command fields because it
only filters out undefined; update the validation in the test that iterates over
loadHooks() / config.hooks so the commandFields array only counts true non-empty
strings: change the filter on [handler.command, handler.bash,
handler.powershell] to accept values where typeof value === "string" and
value.trim() !== "" (reject null, empty, or whitespace-only), then assert
commandFields.length === 1 while leaving the handler.type check unchanged.
---
Nitpick comments:
In `@plugin/scripts/prompt-submit.mjs`:
- Line 25: The fallback for sessionId uses the || operator which treats empty
strings as falsy while prompt uses ??; change the sessionId normalization to use
nullish coalescing (replace the || chain with ??) so it matches the prompt
behavior (i.e., const sessionId = data.session_id ?? data.sessionId ??
"unknown") and scan other normalized fields in this module to ensure all use ??
for consistent null/undefined-only fallbacks.
In `@plugin/scripts/subagent-start.mjs`:
- Around line 26-28: The extraction uses inconsistent fallback operators:
sessionId uses || while agentId and agentType use ??, causing empty strings to
behave differently; update the sessionId assignment (symbol: sessionId) to use
the nullish coalescing operator like agentId and agentType (symbols: agentId,
agentType) — e.g., replace the || fallback with ?? so sessionId =
data.session_id ?? data.sessionId ?? "unknown" to ensure consistent "use if
provided" semantics across all three fields.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 2c0bfbec-f4cc-4ccb-8904-bd231877af9c
📒 Files selected for processing (33)
AGENTS.mdREADME.mdplugin/.mcp.copilot.jsonplugin/hooks/hooks.copilot.jsonplugin/plugin.jsonplugin/scripts/notification.mjsplugin/scripts/post-tool-failure.mjsplugin/scripts/post-tool-use.mjsplugin/scripts/pre-compact.mjsplugin/scripts/pre-tool-use.mjsplugin/scripts/prompt-submit.mjsplugin/scripts/session-end.mjsplugin/scripts/session-start.mjsplugin/scripts/stop.mjsplugin/scripts/subagent-start.mjsplugin/scripts/subagent-stop.mjssrc/cli.tssrc/cli/connect/copilot-cli.tssrc/cli/connect/index.tssrc/cli/connect/util.tssrc/hooks/notification.tssrc/hooks/post-tool-failure.tssrc/hooks/post-tool-use.tssrc/hooks/pre-compact.tssrc/hooks/pre-tool-use.tssrc/hooks/prompt-submit.tssrc/hooks/session-end.tssrc/hooks/session-start.tssrc/hooks/stop.tssrc/hooks/subagent-start.tssrc/hooks/subagent-stop.tstest/cli-connect.test.tstest/copilot-plugin.test.ts
Addresses upstream AI review suggestions by aligning the Copilot preToolUse matcher with the hook allowlist, narrowing hook payload fields at runtime, normalizing subagent fallbacks, and tightening hook config validation. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Contributor
Author
|
Implemented the small CodeRabbit hardening suggestions in |
Contributor
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@src/hooks/pre-tool-use.ts`:
- Line 91: The sessionId assignment currently casts then falls back which can
let non-string truthy values through; update the runtime narrowing in
src/hooks/pre-tool-use.ts to mirror the pattern used for toolName: check typeof
data.session_id === "string" or typeof data.sessionId === "string" and use that
string, otherwise default to "unknown" so only actual strings are sent in the
request; adjust the variable named sessionId accordingly.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: e21b0704-7a4c-4e5e-b590-4cbd4deb570d
📒 Files selected for processing (10)
plugin/hooks/hooks.copilot.jsonplugin/scripts/notification.mjsplugin/scripts/pre-tool-use.mjsplugin/scripts/subagent-start.mjsplugin/scripts/subagent-stop.mjssrc/hooks/notification.tssrc/hooks/pre-tool-use.tssrc/hooks/subagent-start.tssrc/hooks/subagent-stop.tstest/copilot-plugin.test.ts
🚧 Files skipped from review as they are similar to previous changes (9)
- plugin/hooks/hooks.copilot.json
- src/hooks/notification.ts
- plugin/scripts/notification.mjs
- plugin/scripts/subagent-start.mjs
- plugin/scripts/pre-tool-use.mjs
- plugin/scripts/subagent-stop.mjs
- src/hooks/subagent-start.ts
- src/hooks/subagent-stop.ts
- test/copilot-plugin.test.ts
Contributor
Author
|
Note on CodeRabbit's "Docstring Coverage" warning: this repository does not currently enforce docstring/JSDoc coverage in CI, and this PR follows the existing TypeScript style of using descriptive names plus sparse comments for non-obvious behavior. Adding broad docstrings to satisfy a generic 80% threshold would be unrelated churn for this Copilot CLI integration. The actionable review comments were addressed in |
Includes GitHub Copilot CLI in the first-run agent picker and adds a regression test so the Copilot setup path remains discoverable. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Contributor
Author
|
Fixed the first-run onboarding gap: GitHub Copilot CLI is now included in the agent picker, with a regression test covering the native setup options. Commit: Random-Word/agentmemory@71f4c6b |
Detect Copilot CLI environment markers during first-run setup so pressing Enter wires the current agent instead of the historical Claude Code default. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Contributor
Author
|
Follow-up fix: first-run onboarding now detects Copilot CLI via COPILOT_CLI / COPILOT_AGENT_SESSION_ID and default-selects GitHub Copilot CLI instead of Claude Code when setup is launched from Copilot. Commit: Random-Word/agentmemory@11a0200 |
Accept Content-Length framed JSON-RPC messages in addition to the existing newline-delimited transport so Copilot CLI can initialize the standalone MCP server. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Contributor
Author
|
Fixed the Copilot CLI MCP timeout found during local plugin testing. Root cause: Copilot CLI initializes stdio MCP servers with Content-Length framed JSON-RPC, while agentmemory's standalone transport only accepted newline-delimited JSON. The transport now supports both framed and newline-delimited messages, with regression tests and a local framed MCP smoke test. Commit: Random-Word/agentmemory@c21bb06 |
Contributor
There was a problem hiding this comment.
🧹 Nitpick comments (1)
src/mcp/transport.ts (1)
246-255: ⚡ Quick winQueue error handling could stall on unexpected rejection.
If
processLineorwriteResponsethrows unexpectedly,queuebecomes a rejected promise. Subsequent.then()calls chain off the rejected promise, skipping handlers and causing all future messages to be dropped.Inline the catch so the queue recovers:
♻️ Proposed fix
parser = createMessageParser((message) => { - queue = queue.then(() => processLine(message, handler, writeResponse)); - void queue.catch((err) => { + queue = queue.then(() => processLine(message, handler, writeResponse)).catch((err) => { process.stderr.write( `[mcp-transport] request processing failed: ${ err instanceof Error ? err.message : String(err) }\n`, ); }); });🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/mcp/transport.ts` around lines 246 - 255, The current chaining of queue = queue.then(() => processLine(message, handler, writeResponse)); with a separate void queue.catch(...) lets queue become permanently rejected and skip future handlers; instead, ensure each appended task handles its own errors so the shared queue always resolves. Replace the two-step pattern by assigning queue = queue.then(() => processLine(message, handler, writeResponse)).catch(err => { process.stderr.write(`[mcp-transport] request processing failed: ${err instanceof Error ? err.message : String(err)}\n`); }); so any rejection from processLine or writeResponse is caught immediately and the queue is recovered for subsequent messages (refer to parser/createMessageParser, processLine, writeResponse, and the queue variable).
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Nitpick comments:
In `@src/mcp/transport.ts`:
- Around line 246-255: The current chaining of queue = queue.then(() =>
processLine(message, handler, writeResponse)); with a separate void
queue.catch(...) lets queue become permanently rejected and skip future
handlers; instead, ensure each appended task handles its own errors so the
shared queue always resolves. Replace the two-step pattern by assigning queue =
queue.then(() => processLine(message, handler, writeResponse)).catch(err => {
process.stderr.write(`[mcp-transport] request processing failed: ${err
instanceof Error ? err.message : String(err)}\n`); }); so any rejection from
processLine or writeResponse is caught immediately and the queue is recovered
for subsequent messages (refer to parser/createMessageParser, processLine,
writeResponse, and the queue variable).
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 15390489-fd19-4c73-9f5f-6c9ebff01796
📒 Files selected for processing (2)
src/mcp/transport.tstest/mcp-transport.test.ts
Ensures pre-tool-use only forwards string session IDs and falls back to unknown for invalid Copilot payload values, with regression coverage for the generated plugin script. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Resolves conflicts with current main, keeps Copilot CLI support intact, and preserves the Codex hook idempotency fix from main. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Keeps upstream install guidance and retains Copilot CLI in the supported agent list. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
|
I am using github copilot CLI, look forward to seeing this feature. |
|
Looking forward for this to be merged as well |
Owner
|
Thanks @Random-Word — VS Code Copilot uses the standard MCP config block; the connect-table docs in README cover the path. |
Contributor
Author
|
@rohitg00 I'm confused - are you deprecating the hooks + plugin approach for Claude Code, Codex, etc? Are you planning to migrate them to MCP only? The hooks + plugin design seems cleaner than MCP to me. |
Owner
|
@Random-Word — reopening, you were right. My close message was based on a wrong premise (VS Code Copilot ≠ Copilot CLI — those are separate products). Re-audited this PR top to bottom and the hooks + plugin path you took is the architecturally correct one. We're keeping hooks + plugin as the primary integration story across all agents; MCP-only is the fallback for clients that can't load plugins. What I liked on the re-read:
Two small bugs and one rebase before this can land:
Take whichever path is easier for you — rebase + 2 fixes yourself, or pass it back to me and I'll do the rebase + push to a maintainer branch with your commits preserved for attribution. |
…view fixes Merge resolution: - 9 hook source files: composed PR's camelCase payload normalization on top of main's resolveProject() + fire-and-forget pattern. - pre-tool-use: kept both PR's defensive sessionId narrowing and main's explicit project handling (data.project, never cwd-as-project). - src/cli/connect/index.ts: copilotCli alongside cline / continue / zed / droid / warp adapters added on main (ADAPTERS now 17 entries). - test/cli-connect.test.ts: expect 17 adapters, dropped duplicate spec. - plugin/scripts/*.mjs regenerated via npm run build (no hand-edits). Review fixes (per rohitg00#534 audit): - AGENTMEMORY_COPILOT_MCP_BLOCK (src/cli/connect/util.ts) + plugin/.mcp.copilot.json: switched env values to ${VAR:-default} shape so parse-strict clients don't drop the server when the user has no shell export (rohitg00#510 lesson). AGENTMEMORY_TOOLS:all also added. - plugin/plugin.json: version 0.9.21 -> 0.9.22, skill count 4 -> 8 to match plugin/.claude-plugin/plugin.json source of truth. - Tests updated to assert the new env shape.
Owner
|
Pushed the rebase + 2 review fixes to What landed in the merge commit:
Review fixes:
Status: PR is now Thanks for pushing back yesterday; the close was wrong and the hooks + plugin path you built is exactly the architecture we want. |
rohitg00
added a commit
that referenced
this pull request
May 28, 2026
Bumps version across 9 files + adds CHANGELOG entry summarizing the 18 commits since v0.9.22. Highlights: - GitHub Copilot CLI first-class support (#534) — plugin + hooks + MCP with LSP-style Content-Length framing on the standalone stdio transport. - Five new MCP adapters: Warp, Cline, Continue, Zed, Droid (#677); ADAPTERS count 11 → 17. - Three silent DX bugs fixed: graph extraction never fired on session end (#666 / #698), status reported zero memories (#666), consolidation defaulted off even with an LLM provider configured (#612 / #696). - Nine telemetry hooks switched to fire-and-forget so they don't block Claude Code's next-prompt boundary (#573 / #688). - Hook project field now sends repo basename instead of full filesystem path so auto-injected context isn't silently filtered out (#474 / #687). - Local-LLM docs: Ollama / LM Studio / vLLM section added (#671 / #697). Version-bump files: package.json, plugin/.claude-plugin/plugin.json, plugin/plugin.json, plugin/.codex-plugin/plugin.json, packages/mcp/package.json, src/version.ts, src/types.ts, src/functions/export-import.ts, test/export-import.test.ts.
rohitg00
added a commit
that referenced
this pull request
May 28, 2026
* chore(release): v0.9.23 Bumps version across 9 files + adds CHANGELOG entry summarizing the 18 commits since v0.9.22. Highlights: - GitHub Copilot CLI first-class support (#534) — plugin + hooks + MCP with LSP-style Content-Length framing on the standalone stdio transport. - Five new MCP adapters: Warp, Cline, Continue, Zed, Droid (#677); ADAPTERS count 11 → 17. - Three silent DX bugs fixed: graph extraction never fired on session end (#666 / #698), status reported zero memories (#666), consolidation defaulted off even with an LLM provider configured (#612 / #696). - Nine telemetry hooks switched to fire-and-forget so they don't block Claude Code's next-prompt boundary (#573 / #688). - Hook project field now sends repo basename instead of full filesystem path so auto-injected context isn't silently filtered out (#474 / #687). - Local-LLM docs: Ollama / LM Studio / vLLM section added (#671 / #697). Version-bump files: package.json, plugin/.claude-plugin/plugin.json, plugin/plugin.json, plugin/.codex-plugin/plugin.json, packages/mcp/package.json, src/version.ts, src/types.ts, src/functions/export-import.ts, test/export-import.test.ts. * chore(release): add #701 + #709 to v0.9.23 CHANGELOG
3 tasks
rohitg00
added a commit
that referenced
this pull request
May 29, 2026
* chore(website): refresh agents grid + logos for v0.9.23 Agents section was stale relative to the supported agent matrix that v0.9.23 ships: - FEATURED cards bumped from 6 to 7 — adds Copilot CLI (full plugin + hooks + MCP from PR #534). Title updates from "SIX FIRST-PARTY" to "SEVEN NATIVE PLUGINS". - MARQUEE tiles expanded from 10 to 17 — adds Warp, Continue, Zed, Droid (the 4 covered by `npx skills add` from PR #677) plus Qwen Code, Antigravity, and Kiro (the 3 from PR #648). - Logos switched from `github.com/<org>.png` avatars and stale third-party CDNs (freelogovectors.net, exafunction.github.io) to svgl.app brand SVGs where available (Anthropic, GitHub, OpenAI, Cursor, Warp, Continue, Zed, Gemini, Google, Qwen, Windsurf) or the agent's own website favicon where svgl doesn't carry the brand (Factory.ai, Kiro, OpenCode, Cline, Roo, Kilo, Goose, Aider, OpenClaw, Nous Research). Cursor and Windsurf logo paths were specifically broken; the freelogovectors URL was unreliable and Codeium → Cognition acquisition stale-dated the Windsurf path. AgentInstall chip row adds Copilot CLI + Warp alongside the existing Cursor / VS Code / Claude Code / Claude Desktop / Gemini / Codex shortcuts. Universal MCP JSON hint and "show more" button both updated to list the agents we actually support now. * fix(website): use correct svgl.app slugs + add hostnames to remotePatterns Previous commit's logo URLs all 404'd: - svgl.app slugs were wrong (e.g. anthropic.svg vs anthropic_white.svg, cursor.svg vs cursor_dark.svg). svgl exposes themed variants (`_light`/`_dark`/`_white`/`_black`) but bare `<slug>.svg` only exists for a few entries. Verified actual URLs via api.svgl.app for every featured + marquee tile. - Next.js Image `remotePatterns` didn't whitelist svgl.app or the agent-domain favicon hosts (factory.ai, kiro.dev, opencode.ai, cline.bot, etc.), so even valid URLs were rejected before fetch. Now uses dark-bg-appropriate variants throughout (white/dark logos on the black page background). For the few brands not in svgl (continue, kiro, opencode, cline, roo, goose, aider, openclaw, hermes, droid), falls back to the agent's own website favicon — each URL HEAD-probed for 200. Dropped stale remotePatterns: exafunction.github.io, www.freelogovectors.net, block.github.io (replaced with goose.dev). Added: continue.dev, goose.dev. * revert(website): restore github.com avatars for existing agents Only new logos (copilot-cli, warp, continue, zed, droid, antigravity, qwen, kiro) and previously-broken ones (cursor, windsurf) use svgl / own-site URLs. Original github.com/<org>.png avatars for claude-code, codex, openclaw, hermes, claude-desktop, gemini, opencode, cline, roo, kilo, goose, aider restored — they were never broken. next.config remotePatterns trimmed to: svgl.app, www.factory.ai, kiro.dev, continue.dev (the only non-github hosts still in use). * fix(assets): pi logo fill white for dark website bg User-provided SVG uses #09090b near-black; invisible on the agents grid's dark background. Same path, fill swapped to #ffffff.
This was referenced Jun 4, 2026
This was referenced Jul 9, 2026
healdigital
pushed a commit
to healdigital/agentmemory
that referenced
this pull request
Jul 27, 2026
* feat: add Copilot CLI plugin asset slice - plugin/.plugin/plugin.json: Copilot manifest with name/version/skills/mcpServers/hooks refs - plugin/.mcp.copilot.json: MCP server config with type:local, npx, env passthrough, tools:[*] - plugin/hooks/hooks.copilot.json: Copilot hooks (version:1) with 11 supported events and PreToolUse matcher - test/copilot-plugin.test.ts: 11 tests covering manifest, MCP config, and hooks validation Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Add Copilot CLI connect support Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Add GitHub Copilot CLI support Adds Copilot CLI support through a root plugin manifest, Copilot-specific MCP and hook configuration, and a connect adapter for MCP-only setup. Includes Windows-safe Copilot MCP command generation, COPILOT_HOME handling, Copilot hook payload normalization, generated hook scripts, and targeted tests for plugin shape, hook execution, and connect behavior. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Harden Copilot hook handling Addresses upstream AI review suggestions by aligning the Copilot preToolUse matcher with the hook allowlist, narrowing hook payload fields at runtime, normalizing subagent fallbacks, and tightening hook config validation. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Add Copilot to first-run onboarding Includes GitHub Copilot CLI in the first-run agent picker and adds a regression test so the Copilot setup path remains discoverable. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Default onboarding to Copilot inside Copilot CLI Detect Copilot CLI environment markers during first-run setup so pressing Enter wires the current agent instead of the historical Claude Code default. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Support framed stdio MCP transport Accept Content-Length framed JSON-RPC messages in addition to the existing newline-delimited transport so Copilot CLI can initialize the standalone MCP server. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Narrow Copilot pre-tool session ids Ensures pre-tool-use only forwards string session IDs and falls back to unknown for invalid Copilot payload values, with regression coverage for the generated plugin script. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --------- Co-authored-by: Ross Story <rostory@microsoft.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Co-authored-by: Rohit Ghumare <ghumare64@gmail.com>
healdigital
pushed a commit
to healdigital/agentmemory
that referenced
this pull request
Jul 27, 2026
* chore(release): v0.9.23 Bumps version across 9 files + adds CHANGELOG entry summarizing the 18 commits since v0.9.22. Highlights: - GitHub Copilot CLI first-class support (rohitg00#534) — plugin + hooks + MCP with LSP-style Content-Length framing on the standalone stdio transport. - Five new MCP adapters: Warp, Cline, Continue, Zed, Droid (rohitg00#677); ADAPTERS count 11 → 17. - Three silent DX bugs fixed: graph extraction never fired on session end (rohitg00#666 / rohitg00#698), status reported zero memories (rohitg00#666), consolidation defaulted off even with an LLM provider configured (rohitg00#612 / rohitg00#696). - Nine telemetry hooks switched to fire-and-forget so they don't block Claude Code's next-prompt boundary (rohitg00#573 / rohitg00#688). - Hook project field now sends repo basename instead of full filesystem path so auto-injected context isn't silently filtered out (rohitg00#474 / rohitg00#687). - Local-LLM docs: Ollama / LM Studio / vLLM section added (rohitg00#671 / rohitg00#697). Version-bump files: package.json, plugin/.claude-plugin/plugin.json, plugin/plugin.json, plugin/.codex-plugin/plugin.json, packages/mcp/package.json, src/version.ts, src/types.ts, src/functions/export-import.ts, test/export-import.test.ts. * chore(release): add rohitg00#701 + rohitg00#709 to v0.9.23 CHANGELOG
healdigital
pushed a commit
to healdigital/agentmemory
that referenced
this pull request
Jul 27, 2026
* chore(website): refresh agents grid + logos for v0.9.23 Agents section was stale relative to the supported agent matrix that v0.9.23 ships: - FEATURED cards bumped from 6 to 7 — adds Copilot CLI (full plugin + hooks + MCP from PR rohitg00#534). Title updates from "SIX FIRST-PARTY" to "SEVEN NATIVE PLUGINS". - MARQUEE tiles expanded from 10 to 17 — adds Warp, Continue, Zed, Droid (the 4 covered by `npx skills add` from PR rohitg00#677) plus Qwen Code, Antigravity, and Kiro (the 3 from PR rohitg00#648). - Logos switched from `github.com/<org>.png` avatars and stale third-party CDNs (freelogovectors.net, exafunction.github.io) to svgl.app brand SVGs where available (Anthropic, GitHub, OpenAI, Cursor, Warp, Continue, Zed, Gemini, Google, Qwen, Windsurf) or the agent's own website favicon where svgl doesn't carry the brand (Factory.ai, Kiro, OpenCode, Cline, Roo, Kilo, Goose, Aider, OpenClaw, Nous Research). Cursor and Windsurf logo paths were specifically broken; the freelogovectors URL was unreliable and Codeium → Cognition acquisition stale-dated the Windsurf path. AgentInstall chip row adds Copilot CLI + Warp alongside the existing Cursor / VS Code / Claude Code / Claude Desktop / Gemini / Codex shortcuts. Universal MCP JSON hint and "show more" button both updated to list the agents we actually support now. * fix(website): use correct svgl.app slugs + add hostnames to remotePatterns Previous commit's logo URLs all 404'd: - svgl.app slugs were wrong (e.g. anthropic.svg vs anthropic_white.svg, cursor.svg vs cursor_dark.svg). svgl exposes themed variants (`_light`/`_dark`/`_white`/`_black`) but bare `<slug>.svg` only exists for a few entries. Verified actual URLs via api.svgl.app for every featured + marquee tile. - Next.js Image `remotePatterns` didn't whitelist svgl.app or the agent-domain favicon hosts (factory.ai, kiro.dev, opencode.ai, cline.bot, etc.), so even valid URLs were rejected before fetch. Now uses dark-bg-appropriate variants throughout (white/dark logos on the black page background). For the few brands not in svgl (continue, kiro, opencode, cline, roo, goose, aider, openclaw, hermes, droid), falls back to the agent's own website favicon — each URL HEAD-probed for 200. Dropped stale remotePatterns: exafunction.github.io, www.freelogovectors.net, block.github.io (replaced with goose.dev). Added: continue.dev, goose.dev. * revert(website): restore github.com avatars for existing agents Only new logos (copilot-cli, warp, continue, zed, droid, antigravity, qwen, kiro) and previously-broken ones (cursor, windsurf) use svgl / own-site URLs. Original github.com/<org>.png avatars for claude-code, codex, openclaw, hermes, claude-desktop, gemini, opencode, cline, roo, kilo, goose, aider restored — they were never broken. next.config remotePatterns trimmed to: svgl.app, www.factory.ai, kiro.dev, continue.dev (the only non-github hosts still in use). * fix(assets): pi logo fill white for dark website bg User-provided SVG uses #09090b near-black; invisible on the agents grid's dark background. Same path, fill swapped to #ffffff.
This was referenced Jul 31, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds first-class GitHub Copilot CLI support to agentmemory using the same hybrid approach already used for other coding agents:
plugin/plugin.json.agentmemory connect copilot-cli.What changed
Copilot plugin assets
plugin/plugin.json, the root Copilot CLI plugin manifest.plugin/.mcp.copilot.jsonfor Copilot's MCP server config shape.plugin/hooks/hooks.copilot.jsonwith Copilot event names and lowercase tool matchers.agentmemory connect copilot-clisrc/cli/connect/copilot-cli.tsto mergemcpServers.agentmemoryinto Copilot CLI'smcp-config.json.COPILOT_HOME; otherwise defaults to~/.copilot/mcp-config.json.--forceand--dry-run.copilot-clisetup on Windows while preserving existing manual-install behavior for other Windows adapters.cmd.exe /d /s /c npx -y @agentmemory/mcpcommand block for Copilot MCP config on Windows.First-run onboarding
COPILOT_CLI/COPILOT_AGENT_SESSION_IDand default-selectscopilot-cliinstead of the historical Claude Code default when setup is launched from Copilot.Hook compatibility
sessionIdtoolNametoolArgstoolResultuserPrompterrorMessagenotificationTypeplugin/scripts/*.mjsbundles so plugin runtime scripts match the TypeScript sources.MCP transport compatibility
Content-Lengthframed JSON-RPC messages used by Copilot CLI, andContent-Lengthafter receiving framed input, while preserving newline responses for newline clients.Documentation and maintenance rules
AGENTS.mdconsistency rules so future tool-count or version changes include the Copilot plugin manifest/config where relevant.Testing
Validated locally:
npm test -- test/copilot-plugin.test.ts test/cli-connect.test.tsnpm test -- test/onboarding.test.ts test/cli-connect.test.ts test/copilot-plugin.test.tsnpm test -- test/mcp-transport.test.ts test/mcp-standalone-proxy.test.ts test/onboarding.test.ts test/cli-connect.test.ts test/copilot-plugin.test.tsnpx tsdownagentmemoryMCP server and exposed its tools.Validated in GitHub Actions on the fork:
mainCI was green when compared: https://github.com/rohitg00/agentmemory/actions/runs/26066228713Notes:
Review process
This was implemented in parallel worktrees and merged back into
feature/copilot-cli-support:The branch was reviewed with two frontier-model passes:
Validated findings were addressed before the upstream PR was opened. Additional issues discovered during local Copilot CLI testing were fixed and pushed as follow-up commits:
Content-Lengthframed JSON-RPC.