Add shared AgentDB MCP import and wire deep-report for large-scale discussion search#32771
Conversation
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
| @@ -0,0 +1,26 @@ | |||
| --- | |||
| mcp-servers: | |||
There was a problem hiding this comment.
@copilot mount database on cache-memory folder
There was a problem hiding this comment.
Updated AgentDB to store its DB in cache-memory (AGENTDB_PATH: /tmp/gh-aw/cache-memory/discussions.rvf) and regenerated the lockfile in commit 1cdaf8b.
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
| @@ -0,0 +1,26 @@ | |||
| --- | |||
| mcp-servers: | |||
There was a problem hiding this comment.
@copilot configure a custom agentdb cache memory and store file in that path
There was a problem hiding this comment.
Done in 89c127d: AgentDB now uses a dedicated cache-memory path (/tmp/gh-aw/cache-memory/agentdb/) with AGENTDB_PATH pointing to /tmp/gh-aw/cache-memory/agentdb/discussions.rvf, and the workflow now creates that custom cache directory before MCP startup.
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
|
@copilot configure cache-memory with custom id to avoid clashes |
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
Addressed in |
| key: agentdb-${{ env.GH_AW_WORKFLOW_ID_SANITIZED }} | ||
| steps: | ||
| - name: Ensure AgentDB cache path | ||
| run: mkdir -p /tmp/gh-aw/cache-memory/agentdb |
There was a problem hiding this comment.
@copilot ensure this path matches the cache memory path
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
There was a problem hiding this comment.
Pull request overview
Adds a shared AgentDB MCP component and wires it into the deep-report workflow so discussion analysis can use AgentDB-backed semantic/hybrid search.
Changes:
- Added shared AgentDB MCP configuration with cache-memory setup.
- Imported AgentDB into
deep-report.mdand updated the discussion intelligence instructions. - Regenerated
deep-report.lock.ymlto include the AgentDB server and tool permissions.
Show a summary per file
| File | Description |
|---|---|
.github/workflows/shared/mcp/agentdb.md |
Defines the reusable AgentDB MCP server and storage path. |
.github/workflows/deep-report.md |
Imports AgentDB and instructs the report workflow to use it for discussion corpus analysis. |
.github/workflows/deep-report.lock.yml |
Compiled workflow output reflecting the new import and MCP configuration. |
Copilot's findings
Tip
Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
- Files reviewed: 3/3 changed files
- Comments generated: 1
| mcp-servers: | ||
| agentdb: | ||
| command: "npx" | ||
| args: ["agentdb@alpha", "mcp", "start"] |
This change introduces AgentDB as a reusable shared MCP component and uses it in a daily workflow that analyzes Discussions at scale. The deep-report workflow now explicitly routes corpus-level discussion analysis through AgentDB semantic/hybrid search.
Shared MCP component: AgentDB
.github/workflows/shared/mcp/agentdb.md.agentdbMCP server usingnpx agentdb@alpha mcp start.AGENTDB_PATH(/tmp/gh-aw/agentdb/discussions.rvf) for discussion-memory indexing/search in workflow runs.Daily workflow integration
.github/workflows/deep-report.mdimports to includeshared/mcp/agentdb.md.Compiled output alignment
deep-report.lock.ymlto reflect updated frontmatter/import graph.