deepflow 0.1.86 → 0.1.87

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "deepflow",
-  "version": "0.1.86",
+  "version": "0.1.87",
   "description": "Doing reveals what thinking can't predict — spec-driven iterative development for Claude Code",
   "keywords": [
     "claude",

package/src/commands/df/discover.md

@@ -1,7 +1,7 @@
 ---
 name: df:discover
 description: Explore a problem space deeply through structured questioning to surface requirements and constraints
-allowed-tools: [AskUserQuestion, Read]
+allowed-tools: [AskUserQuestion, Read, Agent]
 ---
 
 # /df:discover — Deep Problem Exploration
@@ -10,9 +10,9 @@ allowed-tools: [AskUserQuestion, Read]
 
 You are a Socratic questioner. Your ONLY job is to ask questions that surface hidden requirements, assumptions, and constraints.
 
-**NEVER:** Read source files, use Glob/Grep, spawn agents, create files (except `.deepflow/decisions.md`), run git, use TaskOutput, use Task tool, use EnterPlanMode, use ExitPlanMode
+**NEVER:** Read source files directly, use Glob/Grep directly, proactively spawn agents, create files (except `.deepflow/decisions.md`), run git, use TaskOutput, use Task tool, use EnterPlanMode, use ExitPlanMode
 
-**ONLY:** Ask questions using `AskUserQuestion` tool, respond conversationally
+**ONLY:** Ask questions using `AskUserQuestion` tool, respond conversationally, and spawn context-fetch agents **when the user explicitly requests it**
 
 ---
 
@@ -91,6 +91,46 @@ Example questions:
 - Keep your responses short between questions — don't lecture
 - Acknowledge answers briefly before asking the next question
 
+### On-Demand Context Fetching
+
+When the user explicitly asks you to look at code or a URL (e.g., "olha no código", "vê esse link", "look at src/auth/", "check https://docs.example.com"), fetch context using a sub-agent.
+
+**Trigger:** Intent-based detection — the user must explicitly request it. NEVER proactively fetch context.
+
+**For codebase context:**
+```
+Agent(subagent_type="Explore", model="haiku", prompt="""
+Read and summarize the following: {what the user asked to see}
+
+Rules:
+- Return ONLY factual observations: what files exist, what functions/types are defined, what patterns are used
+- Do NOT suggest solutions, improvements, or architectural changes
+- Do NOT give opinions on code quality
+- Keep response under 4000 tokens
+- Format: bullet points of facts
+""")
+```
+
+**For URL context:**
+```
+Agent(subagent_type="Explore", model="haiku", prompt="""
+Use the browse-fetch skill to fetch this URL: {url}
+
+Then summarize what the page contains.
+
+Rules:
+- Return ONLY factual observations: what the documentation says, what APIs are described, what patterns are shown
+- Do NOT suggest how to use this in the project
+- Do NOT give opinions or recommendations
+- Keep response under 4000 tokens
+- Format: bullet points of facts
+""")
+```
+
+**After receiving context:** Briefly share the factual summary with the user, then **resume Socratic questioning** incorporating the new facts. Do NOT shift to suggesting solutions.
+
+**Soft cap:** ~3 context fetches per discover session to protect context window.
+
 ### When the User Wants to Move On
 When the user signals they want to advance (e.g., "I think that's enough", "let's move on", "ready for next step"):
 

package/src/commands/df/execute.md

@@ -125,7 +125,7 @@ Before spawning: `TaskUpdate(taskId: native_id, status: "in_progress")` — acti
 
 **Token tracking — record start:**
 ```
-start_percentage = !`cat .deepflow/context.json 2>/dev/null | python3 -c "import sys,json; d=json.load(sys.stdin); print(d.get('percentage',''))" 2>/dev/null || echo ''`
+start_percentage = !`grep -o '"percentage":[0-9]*' .deepflow/context.json 2>/dev/null | grep -o '[0-9]*' || echo ''`
 start_timestamp  = !`date -u +%Y-%m-%dT%H:%M:%SZ`
 ```
 Store both values in memory (keyed by task_id) for use after ratchet completes. Omit if context.json unavailable.
@@ -193,32 +193,24 @@ After ratchet checks complete, truncate command output for context efficiency:
 After all checks pass, compute and write the token block to `.deepflow/results/T{N}.yaml`:
 
 ```
-end_percentage = !`cat .deepflow/context.json 2>/dev/null | python3 -c "import sys,json; d=json.load(sys.stdin); print(d.get('percentage',''))" 2>/dev/null || echo ''`
+end_percentage = !`grep -o '"percentage":[0-9]*' .deepflow/context.json 2>/dev/null | grep -o '[0-9]*' || echo ''`
 ```
 
 Parse `.deepflow/token-history.jsonl` to sum token fields for lines whose `timestamp` falls between `start_timestamp` and `end_timestamp` (ISO 8601 compare):
 ```bash
-python3 - <<'EOF'
-import json, sys
-from datetime import datetime, timezone
-
-start = "REPLACE_start_timestamp"
-end   = "REPLACE_end_timestamp"   # current time: date -u +%Y-%m-%dT%H:%M:%SZ
-
-totals = {"input_tokens": 0, "cache_creation_input_tokens": 0, "cache_read_input_tokens": 0}
-try:
-    with open(".deepflow/token-history.jsonl") as f:
-        for line in f:
-            entry = json.loads(line)
-            ts = entry.get("timestamp", "")
-            if start <= ts <= end:
-                for k in totals:
-                    totals[k] += entry.get(k, 0)
-except FileNotFoundError:
-    sys.exit(0)
-
-print(json.dumps(totals))
-EOF
+awk -v start="REPLACE_start_timestamp" -v end="REPLACE_end_timestamp" '
+{
+  ts=""; inp=0; cre=0; rd=0
+  if (match($0, /"timestamp":"[^"]*"/)) { ts=substr($0, RSTART+13, RLENGTH-14) }
+  if (ts >= start && ts <= end) {
+    if (match($0, /"input_tokens":[0-9]+/)) inp=substr($0, RSTART+15, RLENGTH-15)
+    if (match($0, /"cache_creation_input_tokens":[0-9]+/)) cre=substr($0, RSTART+30, RLENGTH-30)
+    if (match($0, /"cache_read_input_tokens":[0-9]+/)) rd=substr($0, RSTART+26, RLENGTH-26)
+    si+=inp; sc+=cre; sr+=rd
+  }
+}
+END { printf "{\"input_tokens\":%d,\"cache_creation_input_tokens\":%d,\"cache_read_input_tokens\":%d}\n", si+0, sc+0, sr+0 }
+' .deepflow/token-history.jsonl 2>/dev/null || echo '{}'
 ```
 
 Append (or create) `.deepflow/results/T{N}.yaml` with the following block. Use shell injection to read the existing file first:
@@ -237,7 +229,7 @@ tokens:
   cache_read_input_tokens: {sum from jsonl}
 ```
 
-**Omit entirely if:** context.json was unavailable at start OR end, OR token-history.jsonl is missing, OR python3 is unavailable. Never fail the ratchet due to token tracking errors.
+**Omit entirely if:** context.json was unavailable at start OR end, OR token-history.jsonl is missing, OR awk is unavailable. Never fail the ratchet due to token tracking errors.
 
 **Evaluate:** All pass + no violations → commit stands. Any failure → attempt partial salvage before reverting:
 

package/src/commands/df/update.md

@@ -5,7 +5,7 @@ description: Update or uninstall deepflow, check installed version
 
 # /df:update — Update deepflow
 
-## Update
+**ACTION REQUIRED:** Immediately run the update command below. Do NOT ask for confirmation — the user already confirmed by running `/df:update`.
 
 ```bash
 npx deepflow@latest
@@ -15,6 +15,8 @@ Auto-detects existing installation and updates it.
 
 ## Uninstall
 
+To uninstall instead, run:
+
 ```bash
 npx deepflow --uninstall
 ```