Add v0.1.0 project plan: 37 beans across 7 epics

Milestone: v0.1.0 - Core Session Inspector
Epics: CLI Foundation, Data Parsing & Models, Database & Caching,
Output Formatting, Sessions Commands, Agent Commands, Stub Commands

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

.beans/claudbg-3wel--v010-core-session-inspector.md

@@ -0,0 +1,10 @@
+---
+# claudbg-3wel
+title: v0.1.0 - Core Session Inspector
+status: todo
+type: milestone
+created_at: 2026-03-27T19:36:45Z
+updated_at: 2026-03-27T19:36:45Z
+---
+
+First release: CLI foundation, session/agent data parsing, DB caching, and core inspect commands (list, dump, transcribe).

.beans/claudbg-4d4h--sessions-commands.md

@@ -0,0 +1,15 @@
+---
+# claudbg-4d4h
+title: Sessions Commands
+status: todo
+type: epic
+priority: normal
+created_at: 2026-03-27T19:36:52Z
+updated_at: 2026-03-27T19:41:00Z
+parent: claudbg-3wel
+blocked_by:
+    - claudbg-6wkk
+    - claudbg-g7zj
+---
+
+sessions list, dump, transcribe subcommands and --follow streaming flag.

.beans/claudbg-6wkk--database-caching.md

@@ -0,0 +1,14 @@
+---
+# claudbg-6wkk
+title: Database & Caching
+status: todo
+type: epic
+priority: normal
+created_at: 2026-03-27T19:36:52Z
+updated_at: 2026-03-27T19:41:00Z
+parent: claudbg-3wel
+blocked_by:
+    - claudbg-mztt
+---
+
+libsql setup, DB schema (normalized + denormalized), lazy sync, and index command.

.beans/claudbg-76fy--implement-jsonl-session-file-reader.md

@@ -0,0 +1,11 @@
+---
+# claudbg-76fy
+title: Implement JSONL session file reader
+status: todo
+type: task
+created_at: 2026-03-27T19:39:15Z
+updated_at: 2026-03-27T19:39:15Z
+parent: claudbg-mztt
+---
+
+Async function that reads a .jsonl file line by line using tokio, deserializes each line into SessionEntry, and skips/logs malformed lines gracefully. Must handle partial last lines (live session writes).

.beans/claudbg-7vkw--index-subcommand-full-resync.md

@@ -0,0 +1,12 @@
+---
+# claudbg-7vkw
+title: 'index subcommand: full resync'
+status: todo
+type: task
+priority: normal
+created_at: 2026-03-27T19:39:33Z
+updated_at: 2026-03-27T19:39:44Z
+parent: claudbg-6wkk
+---
+
+Implement `claudbg index` which walks all discovered session and agent files and syncs them into the DB (respecting lazy-sync mtime/size checks). Add --force flag to clear all cached data and rebuild from scratch. Print progress (N sessions indexed, M skipped).

.beans/claudbg-ag0r--follow-flag-for-agents-commands.md

@@ -0,0 +1,11 @@
+---
+# claudbg-ag0r
+title: follow flag for agents commands
+status: todo
+type: task
+created_at: 2026-03-27T19:40:39Z
+updated_at: 2026-03-27T19:40:39Z
+parent: claudbg-g12t
+---
+
+Add --follow flag to 'agents dump' and 'agents transcribe', mirroring the sessions --follow implementation. Watches the sub-agent JSONL file for new appended lines.

.beans/claudbg-b1pa--stub-commands.md

@@ -0,0 +1,11 @@
+---
+# claudbg-b1pa
+title: Stub Commands
+status: todo
+type: epic
+created_at: 2026-03-27T19:36:52Z
+updated_at: 2026-03-27T19:36:52Z
+parent: claudbg-3wel
+---
+
+tui and query subcommands that print 'coming soon\!' placeholders.

.beans/claudbg-bh11--sessions-list-command.md

@@ -0,0 +1,11 @@
+---
+# claudbg-bh11
+title: sessions list command
+status: todo
+type: task
+created_at: 2026-03-27T19:40:02Z
+updated_at: 2026-03-27T19:40:02Z
+parent: claudbg-4d4h
+---
+
+Implement 'claudbg sessions list'. Default columns: session ID (short 8-char), date (ISO or human-relative), project path, model, message count. With --verbose: full UUID. Supports --output flag. Sorted chronologically (most recent first). Lazily syncs session index before querying.

.beans/claudbg-bt5p--xml-output-renderer.md

@@ -0,0 +1,11 @@
+---
+# claudbg-bt5p
+title: XML output renderer
+status: todo
+type: task
+created_at: 2026-03-27T19:40:53Z
+updated_at: 2026-03-27T19:40:53Z
+parent: claudbg-g7zj
+---
+
+Implement --output=xml renderer. Not a strict XML schema — use XML tags as section guards to aid model attention (e.g. <session>, <message role='assistant'>, <tool_use name='Bash'>, <thinking>, <content>). Mechanically wraps each logical section. No DTD or schema validation required.

.beans/claudbg-d3aj--session-statistics-computation.md

@@ -0,0 +1,11 @@
+---
+# claudbg-d3aj
+title: Session statistics computation
+status: todo
+type: task
+created_at: 2026-03-27T19:39:15Z
+updated_at: 2026-03-27T19:39:15Z
+parent: claudbg-mztt
+---
+
+Given a parsed list of SessionEntries, compute: total input/output/cache tokens (sum across all assistant message usage fields), tool call counts keyed by tool name, session duration (sum of durationMs from system messages with subtype=turn_duration), model (from first assistant message). Return a SessionStats struct.

.beans/claudbg-edmr--lazy-sync-process-session-files-on-read.md

@@ -0,0 +1,11 @@
+---
+# claudbg-edmr
+title: 'Lazy sync: process session files on read'
+status: todo
+type: task
+created_at: 2026-03-27T19:39:33Z
+updated_at: 2026-03-27T19:39:33Z
+parent: claudbg-6wkk
+---
+
+Before serving any session data, check if the session's file_mtime and file_size in the DB match the file on disk. If not (or if absent), parse the file and upsert into all tables. Skip files that are already up to date. This runs transparently before every command that reads session data.

.beans/claudbg-ei6j--sessions-dump-command.md

@@ -0,0 +1,11 @@
+---
+# claudbg-ei6j
+title: sessions dump command
+status: todo
+type: task
+created_at: 2026-03-27T19:40:02Z
+updated_at: 2026-03-27T19:40:02Z
+parent: claudbg-4d4h
+---
+
+Implement 'claudbg sessions dump <session-id>'. Minimal processing: read each JSONL line and render it as a table row with columns: #, timestamp, type, role, content preview (truncated). Expands newlines in content for readability. With --verbose: show full content, no truncation. Accepts short (8-char prefix) or full session ID.

.beans/claudbg-fkyt--json-output-renderer.md

@@ -0,0 +1,11 @@
+---
+# claudbg-fkyt
+title: JSON output renderer
+status: todo
+type: task
+created_at: 2026-03-27T19:40:52Z
+updated_at: 2026-03-27T19:40:52Z
+parent: claudbg-g7zj
+---
+
+Implement --output=json renderer: serialize the data model to pretty-printed JSON using serde_json. All commands should produce clean, parseable JSON with no extra decorations.

.beans/claudbg-ftb9--table-output-renderer.md

@@ -0,0 +1,11 @@
+---
+# claudbg-ftb9
+title: Table output renderer
+status: todo
+type: task
+created_at: 2026-03-27T19:40:52Z
+updated_at: 2026-03-27T19:40:52Z
+parent: claudbg-g7zj
+---
+
+Implement the table renderer using comfy-table or tabled. Define a trait (e.g. Renderable) that all data types implement to produce a table. Used as the default --output=table format. Handle terminal width gracefully (truncate or wrap columns).

.beans/claudbg-g12t--agent-commands.md

@@ -0,0 +1,15 @@
+---
+# claudbg-g12t
+title: Agent Commands
+status: todo
+type: epic
+priority: normal
+created_at: 2026-03-27T19:36:52Z
+updated_at: 2026-03-27T19:41:00Z
+parent: claudbg-3wel
+blocked_by:
+    - claudbg-6wkk
+    - claudbg-g7zj
+---
+
+agents list, dump, transcribe subcommands for sub-agent sessions and --follow flag.

.beans/claudbg-g5uv--session-file-discovery.md

@@ -0,0 +1,11 @@
+---
+# claudbg-g5uv
+title: Session file discovery
+status: todo
+type: task
+created_at: 2026-03-27T19:39:15Z
+updated_at: 2026-03-27T19:39:15Z
+parent: claudbg-mztt
+---
+
+Walk ~/.claude/projects/ to enumerate all session .jsonl files. Decode URL-encoded directory names (e.g. -run-media-pop-... → /run/media/pop/...) to recover the project path. Return a list of SessionRef { session_id: Uuid, project_path: PathBuf, file_path: PathBuf, modified_at: SystemTime }.

.beans/claudbg-g7zj--output-formatting.md

@@ -0,0 +1,14 @@
+---
+# claudbg-g7zj
+title: Output Formatting
+status: todo
+type: epic
+priority: normal
+created_at: 2026-03-27T19:36:52Z
+updated_at: 2026-03-27T19:41:00Z
+parent: claudbg-3wel
+blocked_by:
+    - claudbg-h7xu
+---
+
+Table (default), JSON, and XML output renderers behind the global --output flag.

.beans/claudbg-goc1--libsql-db-setup-and-connection.md

@@ -0,0 +1,11 @@
+---
+# claudbg-goc1
+title: libsql DB setup and connection
+status: todo
+type: task
+created_at: 2026-03-27T19:39:32Z
+updated_at: 2026-03-27T19:39:32Z
+parent: claudbg-6wkk
+---
+
+Initialize a libsql connection to ~/.claude/claudbg.db using tokio. Create the DB file if it does not exist. Expose a DbPool or Connection handle shared across the app. Support a --clear flag that drops and recreates the DB.

.beans/claudbg-h7xu--cli-foundation.md

@@ -0,0 +1,11 @@
+---
+# claudbg-h7xu
+title: CLI Foundation
+status: todo
+type: epic
+created_at: 2026-03-27T19:36:52Z
+updated_at: 2026-03-27T19:36:52Z
+parent: claudbg-3wel
+---
+
+Main CLI entry point, clap subcommand structure, global flags, output format infrastructure, and shared utilities.

.beans/claudbg-hskl--denormalized-raw-sessions-table.md

@@ -0,0 +1,11 @@
+---
+# claudbg-hskl
+title: Denormalized raw sessions table
+status: todo
+type: task
+created_at: 2026-03-27T19:39:32Z
+updated_at: 2026-03-27T19:39:32Z
+parent: claudbg-6wkk
+---
+
+Create a raw_sessions table (session_id TEXT PK, raw_jsonl TEXT, file_mtime INTEGER, file_size INTEGER) storing the full file content as a blob. Used for dump and fast retrieval when normalized tables are overkill.

.beans/claudbg-i09w--app-error-type-and-result-alias.md

@@ -0,0 +1,11 @@
+---
+# claudbg-i09w
+title: App error type and Result alias
+status: todo
+type: task
+created_at: 2026-03-27T19:38:56Z
+updated_at: 2026-03-27T19:38:56Z
+parent: claudbg-h7xu
+---
+
+Define a top-level AppError enum (IoError, ParseError, DbError, NotFound, etc.) and a Result<T> alias. Use thiserror or a similar crate.

.beans/claudbg-i8tm--query-subcommand-stub.md

@@ -0,0 +1,11 @@
+---
+# claudbg-i8tm
+title: query subcommand stub
+status: todo
+type: task
+created_at: 2026-03-27T19:40:53Z
+updated_at: 2026-03-27T19:40:53Z
+parent: claudbg-b1pa
+---
+
+Add 'claudbg query' subcommand via clap. Handler prints: 'query: coming soon!' and exits 0. Include placeholder help text describing the intended behavior: filter/search sessions by project, date, tool use, with key:value and key:glob syntax.

.beans/claudbg-iwqj--agents-transcribe-command.md

@@ -0,0 +1,11 @@
+---
+# claudbg-iwqj
+title: agents transcribe command
+status: todo
+type: task
+created_at: 2026-03-27T19:40:39Z
+updated_at: 2026-03-27T19:40:39Z
+parent: claudbg-g12t
+---
+
+Implement 'claudbg agents transcribe <session-id> <agent-id>'. Same structure as sessions transcribe: stats header (token counts, tool use breakdown, duration, model) + pretty-printed chat log. Respects --include flag.

.beans/claudbg-jupi--sub-agent-file-discovery.md

@@ -0,0 +1,11 @@
+---
+# claudbg-jupi
+title: Sub-agent file discovery
+status: todo
+type: task
+created_at: 2026-03-27T19:39:15Z
+updated_at: 2026-03-27T19:39:15Z
+parent: claudbg-mztt
+---
+
+Given a session ID, locate its subagents/ directory under ~/.claude/projects/{project-dir}/{session-id}/subagents/. Enumerate agent-{id}.jsonl files and their corresponding .meta.json files. Return AgentRef { agent_id: String, agent_type: Option<String>, session_id: Uuid, file_path: PathBuf }.

.beans/claudbg-ki43--agents-list-command.md

@@ -0,0 +1,11 @@
+---
+# claudbg-ki43
+title: agents list command
+status: todo
+type: task
+created_at: 2026-03-27T19:40:39Z
+updated_at: 2026-03-27T19:40:39Z
+parent: claudbg-g12t
+---
+
+Implement 'claudbg agents list <session-id>'. Columns: agent ID (short 8-char hex), agent type (from .meta.json, e.g. Explore/Plan/general-purpose), message count. With --verbose: full agent ID. Accepts short or full session ID. Supports --output flag.

.beans/claudbg-kzcs--agents-dump-command.md

@@ -0,0 +1,11 @@
+---
+# claudbg-kzcs
+title: agents dump command
+status: todo
+type: task
+created_at: 2026-03-27T19:40:39Z
+updated_at: 2026-03-27T19:40:39Z
+parent: claudbg-g12t
+---
+
+Implement 'claudbg agents dump <session-id> <agent-id>'. Same as sessions dump but reads the sub-agent JSONL file. Minimal processing: table row per entry with #, timestamp, type, role, content preview. --verbose for full content.

.beans/claudbg-lmdl--tui-subcommand-stub.md

@@ -0,0 +1,11 @@
+---
+# claudbg-lmdl
+title: tui subcommand stub
+status: todo
+type: task
+created_at: 2026-03-27T19:40:53Z
+updated_at: 2026-03-27T19:40:53Z
+parent: claudbg-b1pa
+---
+
+Add 'claudbg tui' subcommand via clap. Handler prints: 'tui: coming soon!' and exits 0. Subcommand should have a brief description in --help output.

.beans/claudbg-m4p1--sessions-transcribe-command.md

@@ -0,0 +1,11 @@
+---
+# claudbg-m4p1
+title: sessions transcribe command
+status: todo
+type: task
+created_at: 2026-03-27T19:40:02Z
+updated_at: 2026-03-27T19:40:02Z
+parent: claudbg-4d4h
+---
+
+Implement 'claudbg sessions transcribe <session-id>'. Header section: SessionStats (token counts, tool call breakdown by name, duration, model). Body: each message rendered as a chat log entry — user text, assistant text blocks, tool use (name + input always shown; output shown only if 'output' is in --include), thinking blocks shown only if 'thinking' in --include. Default --include: tools.

.beans/claudbg-mztt--data-parsing-models.md

@@ -0,0 +1,14 @@
+---
+# claudbg-mztt
+title: Data Parsing & Models
+status: todo
+type: epic
+priority: normal
+created_at: 2026-03-27T19:36:52Z
+updated_at: 2026-03-27T19:41:00Z
+parent: claudbg-3wel
+blocked_by:
+    - claudbg-h7xu
+---
+
+Rust types for all JSONL message variants, file discovery, JSONL reader, and session statistics computation.

.beans/claudbg-nmuu--define-global-cli-flags.md

@@ -0,0 +1,11 @@
+---
+# claudbg-nmuu
+title: Define global CLI flags
+status: todo
+type: task
+created_at: 2026-03-27T19:38:56Z
+updated_at: 2026-03-27T19:38:56Z
+parent: claudbg-h7xu
+---
+
+Add global flags available to all subcommands: --output [table|json|xml] (default: table), --verbose (bool), --include <list> (comma-separated: thinking,output,tools — default: tools). These must be accessible from every command handler.

.beans/claudbg-pkyj--implement-main-cli-entry-point-and-subcommand-stru.md

@@ -0,0 +1,11 @@
+---
+# claudbg-pkyj
+title: Implement main CLI entry point and subcommand structure
+status: todo
+type: task
+created_at: 2026-03-27T19:38:56Z
+updated_at: 2026-03-27T19:38:56Z
+parent: claudbg-h7xu
+---
+
+Wire up clap App with top-level subcommands: sessions, agents, index, tui, query. Each subcommand is a separate clap subcommand group. Binary entry point dispatches to the correct handler.

.beans/claudbg-uls1--define-rust-types-for-jsonl-schema.md

@@ -0,0 +1,11 @@
+---
+# claudbg-uls1
+title: Define Rust types for JSONL schema
+status: todo
+type: task
+created_at: 2026-03-27T19:39:15Z
+updated_at: 2026-03-27T19:39:15Z
+parent: claudbg-mztt
+---
+
+Model the full JSONL schema as serde-deserializable Rust types: SessionEntry enum (user/assistant/system/progress/file-history-snapshot/queue-operation), AssistantContent enum (text/thinking/tool_use), UserContent enum (text/tool_result/image), Usage struct, ToolUseInput, SystemMessage subtypes, ProgressData. All fields should match the on-disk schema discovered from ~/.claude/projects/.

.beans/claudbg-umrv--normalized-db-schema-sessions-messages-tool-uses-t.md

@@ -0,0 +1,11 @@
+---
+# claudbg-umrv
+title: 'Normalized DB schema: sessions, messages, tool_uses tables'
+status: todo
+type: task
+created_at: 2026-03-27T19:39:32Z
+updated_at: 2026-03-27T19:39:32Z
+parent: claudbg-6wkk
+---
+
+Create tables: sessions (session_id, project_path, file_path, first_message_at, last_message_at, model, message_count, file_mtime, file_size), messages (id, session_id, uuid, timestamp, type, role, content_json), tool_uses (id, session_id, message_uuid, tool_name, input_json, result_json). Include indexes on session_id and timestamp.

.beans/claudbg-x7wb--short-uuid-display-utility.md

@@ -0,0 +1,11 @@
+---
+# claudbg-x7wb
+title: Short UUID display utility
+status: todo
+type: task
+created_at: 2026-03-27T19:38:56Z
+updated_at: 2026-03-27T19:38:56Z
+parent: claudbg-h7xu
+---
+
+Utility function: given a UUID string, return the first 8 characters as the short form. Used in sessions list, agents list, etc. when --verbose is not passed.

.beans/claudbg-yugb--add-cargotoml-dependencies.md

@@ -0,0 +1,11 @@
+---
+# claudbg-yugb
+title: Add Cargo.toml dependencies
+status: todo
+type: task
+created_at: 2026-03-27T19:38:56Z
+updated_at: 2026-03-27T19:38:56Z
+parent: claudbg-h7xu
+---
+
+Add clap (derive feature), tokio (full), serde + serde_json, libsql, chrono, uuid, and a table rendering crate (comfy-table or tabled) to Cargo.toml.

.beans/claudbg-zwoj--follow-flag-for-sessions-commands.md

@@ -0,0 +1,11 @@
+---
+# claudbg-zwoj
+title: follow flag for sessions commands
+status: todo
+type: task
+created_at: 2026-03-27T19:40:28Z
+updated_at: 2026-03-27T19:40:28Z
+parent: claudbg-4d4h
+---
+
+Add --follow flag to 'sessions dump' and 'sessions transcribe'. After reading to end of file, continue watching the file for new lines (similar to tail -f). New entries are rendered and appended to output as they arrive. Use tokio file watching or polling.

flake.nix

@@ -67,7 +67,9 @@
             cargo-watch
             cargo-edit
             cargo-nextest
+            clang
             beans
+            jq
           ];
         };
       });