fix: read the progress trail from .datasqrl/ and ignore .datasqrl/ in the deploy check

.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "datasqrl",
   "displayName": "DataSQRL",
-  "version": "0.3.2",
+  "version": "0.3.3",
   "description": "Plan and implement DataSQRL data pipelines with the Dockerized DataSQRL Code Agent.",
   "author": {
     "name": "DataSQRL",

.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "datasqrl",
-  "version": "0.3.2",
+  "version": "0.3.3",
   "description": "Plan and implement DataSQRL data pipelines with the Dockerized DataSQRL Code Agent.",
   "author": {
     "name": "DataSQRL",

.cursor-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "datasqrl",
-  "version": "0.3.2",
+  "version": "0.3.3",
   "description": "Plan and implement DataSQRL data pipelines with the Dockerized DataSQRL Code Agent.",
   "author": {
     "name": "DataSQRL",

DATA_HANDLING.md

@@ -26,7 +26,8 @@ Your coding agent (Claude Code, Codex, Cursor or Copilot) has its own data polic
 The project directory is the only place a run can write. A run writes these files there:
 
 - `adr/requirements_<ts>.md` and `adr/plan_<ts>.md`, the requirements and the plan.
-- `AGENTS.md` and `.claude/`, the agent's instructions, its progress trail and its structured run log (`.claude/codeagent_logs.jsonl.bak`).
+- `AGENTS.md`, the agent's instructions.
+- `.datasqrl/`, the agent's run files: its progress trail, its structured run log (`.datasqrl/datasqrl-agent-logs.jsonl.bak`), its context block and its survey of the repository.
 - `build/`, the compiler output.
 - `.datasqrl_agent_results.json`, the run history.
 - The DataSQRL project files the agent creates or edits.

scripts/datasqrl-agent.sh

@@ -662,7 +662,7 @@ RUN_NAME="datasqrl-agent-${PROJECT_SLUG}-$(_short_hash "$PROJECT_ABS")"
 # PROGRESS_PATH is the human-readable trail the container renders from its log: one record per
 # line, `+HH:MM:SS  KIND   text`, continuation lines indented. Truncated when a run starts and left
 # in place afterwards, so the last run's trail can be revisited until the next run overwrites it.
-PROGRESS_PATH="$PROJECT_ABS/.claude/datasqrl-agent-progress.txt"
+PROGRESS_PATH="$PROJECT_ABS/.datasqrl/datasqrl-agent-progress.txt"
 RESULTS_PATH="$PROJECT_ABS/.datasqrl_agent_results.json"
 
 # --- Shared helpers ----------------------------------------------------------------------------

skills/deploy/SKILL.md

@@ -60,13 +60,12 @@ the working tree.
 Clean tree → deploy it, and say nothing about git.
 
 Dirty tree → find out whether the dirt is work the deployment would carry. Only the files the
-DataSQRL compiler treats as project material matter; `adr/` and `.claude/` are the agent's own
-bookkeeping and are dirty after every run, so counting them would raise a question on every
+DataSQRL compiler treats as project material matter; `adr/`, and `.datasqrl/` are the agent's own bookkeeping and are dirty after every run, so counting them would raise a question on every
 deploy:
 
 ```bash
 git status --porcelain \
-  | grep -vE '^.. (adr/|\.claude/|snapshots/|build/)' \
+  | grep -vE '^.. (adr/|\.datasqrl/|\.claude/|snapshots/|build/)' \
   | grep -E '\.(sqrl|graphqls?|csv|tsv|json|jsonl|ndjson|parquet|orc|avro)$|run-tests\.sh$'
 ```