feat(sleep): add DeepSeek Harness transcript source

docs/reference/cli.md

@@ -134,7 +134,7 @@ Common options for the nightly actions include:
 |---|---|
 | `--project PATH` | Project used for transcript scope, targets, state, and staging (default: current directory) |
 | `--scope invoked\|all` | Harvest this project or all projects |
-| `--source claude\|codex\|copilot\|cursor\|pi\|opencode\|auto` | Transcript source; `auto` keeps Codex-then-Claude precedence and does not select Copilot, Cursor, Pi, or OpenCode |
+| `--source claude\|codex\|copilot\|cursor\|pi\|opencode\|dsh\|auto` | Transcript source; `auto` keeps Codex-then-Claude precedence and does not select Copilot, Cursor, Pi, OpenCode, or DSH |
 | `--backend mock\|claude\|codex\|copilot\|cursor\|pi\|opencode\|handoff\|azure_openai` | Replay/optimizer backend |
 | `--model NAME` | Backend-specific model override |
 | `--cursor-home PATH` | Override `~/.cursor` for Cursor transcript harvesting |
@@ -144,6 +144,7 @@ Common options for the nightly actions include:
 | `--pi-path PATH` | Path to the installed Pi coding-agent CLI |
 | `--opencode-path PATH` | Path to the installed OpenCode CLI |
 | `--opencode-db PATH` | Path to the OpenCode SQLite history database |
+| `--dsh-session-root PATH` | Override the DSH JSONL session root for `--source dsh` (default: `$DSH_HOME/sessions`, or `~/.dsh/sessions`) |
 | `--opencode-tool-replay` | Enable OpenCode tool-aware replay for `tool_called` checks in rule judges |
 | `--preferences TEXT` | House rules supplied to reflection |
 | `--lookback-hours N` | Initial transcript lookback; `0` scans all history |

docs/sleep/README.md

@@ -190,6 +190,48 @@ The managed scheduler records the backend but does not preserve `--source`,
 `~/.skillopt-sleep/config.json`. Use an absolute `pi_path` and verify the
 scheduled account's Pi authentication.
 
+### DeepSeek Harness (DSH)
+
+Use `--source dsh` to read local DSH JSONL sessions. By default, SkillOpt uses
+DSH's standard `$DSH_HOME/sessions` directory, or `~/.dsh/sessions` when
+`DSH_HOME` is unset. Install the optional Zstandard reader first:
+
+```bash
+python -m pip install -e ".[dsh]"
+skillopt-sleep harvest --project "$(pwd)" --source dsh --progress
+```
+
+`--dsh-session-root PATH` remains available only to override that default.
+
+The source reads `session.jsonl` and the default compressed
+`session.jsonl.zstd` files below DSH's project/session directory layout. It
+does not start DSH, require DSH login, connect to a model provider, or modify
+the stored logs. `--source auto` retains Codex-then-Claude precedence and does
+not select DSH.
+
+DSH harvesting keeps human user text, visible assistant text, short tool names,
+timestamps, and positive/negative feedback signals derived from the immutable
+`feedback/record` event. It excludes reasoning, tool arguments and results,
+request/provider metadata, attachments, feedback remarks themselves, and
+injected user-role context. Malformed sessions are silently skipped as a whole;
+other sessions continue. Ordinary fork sessions are retained, while sessions
+explicitly marked as subagents and SkillOpt replay sessions are excluded.
+
+The source follows the current observed DSH session format rather than a fixed
+application-version matrix. A session whose format or event structure cannot be
+safely understood is skipped. Plugin integration and DSH execution are outside
+this source's scope.
+
+The managed scheduler does not preserve `--source` or `--dsh-session-root`.
+Before scheduling, set the source in `~/.skillopt-sleep/config.json`; add
+`dsh_session_root` only when overriding DSH's normal session directory:
+
+```json
+{
+  "transcript_source": "dsh"
+}
+```
+
 ### OpenCode
 
 Use `--source opencode` to read local OpenCode SQLite history without launching

pyproject.toml

@@ -55,6 +55,8 @@ docs = ["mkdocs-material>=9.5.0", "mkdocstrings[python]>=0.24.0"]
 webui = ["gradio>=5.50.0,<7"]
 # Development tools
 dev = ["ruff>=0.4.0", "pytest>=8.0.0"]
+# DeepSeek Harness JSONL transcript source
+dsh = ["zstandard>=0.22.0"]
 # All optional dependencies (except docs/dev/webui)
 all = [
     "alfworld>=0.4.0",

skillopt_sleep/__main__.py

@@ -15,10 +15,11 @@
     --target-skill-path PATH explicit live SKILL.md to stage/adopt
     --tasks-file PATH   reviewed TaskRecord JSON file to replay instead of harvesting
     --backend mock|claude|codex|copilot|cursor|pi|opencode|handoff|azure_openai
-    --source claude|codex|copilot|copilot_cli|cursor|pi|opencode|auto
+    --source claude|codex|copilot|copilot_cli|cursor|pi|opencode|dsh|auto
     --vscode-workspace-storage PATH
     --copilot-cli-session-store PATH
     --opencode-db PATH
+    --dsh-session-root PATH
     --model NAME
     --lookback-hours N
     --auto-adopt
@@ -115,14 +116,16 @@ def _add_common(p: argparse.ArgumentParser) -> None:
     p.add_argument("--cursor-home", default="", help="override ~/.cursor for Cursor session harvest")
     p.add_argument("--pi-home", default="", help="override ~/.pi for Pi session harvest")
     p.add_argument("--source", default="",
-                   choices=["", "claude", "codex", "copilot", "copilot_cli", "cursor", "pi", "opencode", "auto"],
+                   choices=["", "claude", "codex", "copilot", "copilot_cli", "cursor", "pi", "opencode", "dsh", "auto"],
                    help="session transcript source")
     p.add_argument("--vscode-workspace-storage", default="",
                    help="override VS Code User/workspaceStorage root for copilot source")
     p.add_argument("--copilot-cli-session-store", default="",
                    help="override ~/.copilot/session-store.db for copilot_cli source")
     p.add_argument("--opencode-db", default="",
                    help="override the local OpenCode transcript database")
+    p.add_argument("--dsh-session-root", default="",
+                   help="override DSH session root (default: $DSH_HOME/sessions or ~/.dsh/sessions)")
     p.add_argument("--lookback-hours", type=int, default=None,
                    help="harvest window in hours; 0 = scan full history")
     p.add_argument("--edit-budget", type=int, default=0)
@@ -194,6 +197,8 @@ def _cfg_from_args(args, task_meta: Dict[str, Any] | None = None) -> Any:
             if args.opencode_db == ":memory:"
             else os.path.abspath(os.path.expanduser(args.opencode_db))
         )
+    if getattr(args, "dsh_session_root", ""):
+        overrides["dsh_session_root"] = os.path.abspath(os.path.expanduser(args.dsh_session_root))
     lh = getattr(args, "lookback_hours", None)
     if lh is not None:  # --lookback-hours was explicitly passed (0 = full history)
         overrides["lookback_hours"] = lh

skillopt_sleep/config.py

@@ -32,7 +32,8 @@
     "vscode_workspace_storage": "",  # "" => auto-detect platform defaults
     "copilot_cli_session_store": "",  # "" => ~/.copilot/session-store.db
     "opencode_db": "",  # "" => OPENCODE_DB or the OpenCode XDG data path
-    # Explicit sources also include copilot, copilot_cli, cursor, pi, and opencode.
+    "dsh_session_root": "",  # "" => $DSH_HOME/sessions, or ~/.dsh/sessions
+    # Explicit sources also include copilot, copilot_cli, cursor, pi, opencode, and dsh.
     # ``auto`` keeps the established Codex-then-Claude precedence.
     "transcript_source": "claude",
     "projects": "invoked",        # "invoked" | "all" | [list of abs paths]
@@ -164,6 +165,18 @@ def opencode_db_path(self) -> str:
             return ":memory:"
         return os.path.abspath(os.path.expanduser(str(value)))
 
+    @property
+    def dsh_session_root(self) -> str:
+        value = self.data.get("dsh_session_root", "") or ""
+        if value:
+            return os.path.abspath(os.path.expanduser(str(value)))
+        # Match the DSH base bundle: its JSONL persistence root is
+        # dshHomePath("sessions"), where DSH_HOME defaults to ~/.dsh.
+        dsh_home = str(os.environ.get("DSH_HOME", ""))
+        if not dsh_home.strip():
+            dsh_home = os.path.join(os.path.expanduser("~"), ".dsh")
+        return os.path.abspath(os.path.expanduser(os.path.join(dsh_home, "sessions")))
+
     @property
     def vscode_workspace_storage(self) -> str:
         value = self.data.get("vscode_workspace_storage", "") or ""