feat(identity): read claudetools_root from identity.json

- Updated sync.sh to read claudetools_root from identity.json - Updated syncro.md skill to use identity.json for repo path - Updated CLAUDE.md onboarding to include claudetools_root field - Eliminates cross-architecture path detection issues - Fallback to git rev-parse for legacy machines Each machine sets claudetools_root during onboarding, just like vault_path.
2026-05-26 18:44:38 -07:00
parent 7513f21e00
commit f94849fc00
3 changed files with 105 additions and 20 deletions
--- a/.claude/CLAUDE.md
+++ b/.claude/CLAUDE.md
@@ -16,10 +16,11 @@ This repo is shared across multiple team members. **At every session start, BEFO
       "email": "mike@azcomputerguru.com",
       "role": "admin",
       "machine": "<HOSTNAME>",
-       "vault_path": "<absolute path to vault repo on this machine>"
+       "vault_path": "<absolute path to vault repo on this machine>",
+       "claudetools_root": "<absolute path to ClaudeTools repo on this machine>"
     }
     ```
-     Ask the user where the vault repo is cloned on this machine (e.g., `D:/vault`, `~/vault`, `/Users/howard/vault`).
+     Ask the user where the vault repo is cloned (e.g., `D:/vault`, `~/vault`, `/Users/howard/vault`) and where ClaudeTools is cloned (e.g., `D:/claudetools`, `~/ClaudeTools`, `/Users/mike/ClaudeTools`).
   - Set local git config: `git config user.name "<full_name>"` and `git config user.email "<email>"`
   - Set git remote (read `gitea_username` from users.json): `git remote set-url origin https://<gitea_username>@git.azcomputerguru.com/azcomputerguru/claudetools.git`
   - Add hostname to user's `known_machines` in users.json and commit.
--- a/.claude/commands/syncro.md
+++ b/.claude/commands/syncro.md
@@ -76,6 +76,8 @@ If any check fails, complete the missing step before reporting done. This rule f

 When invoked, use the Syncro REST API via `curl`. All requests include `?api_key=<key>` as query parameter (NOT in header — Syncro uses query param auth).

+**Repo root resolution:** All scripts read `claudetools_root` from `.claude/identity.json` (set during machine onboarding). This eliminates cross-architecture path issues. The identity.json file is machine-specific (gitignored) and contains the absolute path to the ClaudeTools repo on each machine.
+
 ### Attribution rule (CRITICAL)

 Every Syncro API call is attributed to the **owner of the API key**. Comments, line items, timer entries, and invoices created by the API are logged as the API user — regardless of who is running the command. So the skill MUST use a per-user API key that matches the actual tech running it, or comments will be misattributed.
@@ -92,8 +94,36 @@ Keys are baked into the skill below. To add a new user: generate a token in Sync
 ```bash
 BASE="https://computerguru.syncromsp.com/api/v1"

+# Get repo root from identity.json (set during machine onboarding)
+# Fallback to dynamic detection for legacy machines that haven't updated identity.json yet
+IDENTITY_PATH="${HOME}/.claude/identity.json"
+if [ ! -f "$IDENTITY_PATH" ]; then
+    # Try in-repo identity.json (gitignored, machine-specific)
+    REPO_ROOT=$(git rev-parse --show-toplevel 2>/dev/null)
+    if [ -n "$REPO_ROOT" ]; then
+        IDENTITY_PATH="$REPO_ROOT/.claude/identity.json"
+    fi
+fi
+
+if [ ! -f "$IDENTITY_PATH" ]; then
+    echo "[ERROR] Cannot locate identity.json - run onboarding first" >&2
+    exit 1
+fi
+
+REPO_ROOT=$(jq -r '.claudetools_root // empty' "$IDENTITY_PATH")
+if [ -z "$REPO_ROOT" ]; then
+    # Legacy fallback for machines without claudetools_root in identity.json
+    REPO_ROOT=$(git rev-parse --show-toplevel 2>/dev/null)
+    if [ -z "$REPO_ROOT" ]; then
+        echo "[ERROR] claudetools_root not set in identity.json and not in a git directory" >&2
+        echo "[ERROR] Add 'claudetools_root' field to $IDENTITY_PATH" >&2
+        exit 1
+    fi
+    echo "[WARNING] Using git-detected repo root. Add 'claudetools_root' to identity.json to avoid this." >&2
+fi
+
 # Per-user keys — actions in Syncro are attributed to the key owner
-USER_ID=$(jq -r '.user // empty' "$CLAUDETOOLS_ROOT/.claude/identity.json")
+USER_ID=$(jq -r '.user // empty' "$IDENTITY_PATH")
 case "$USER_ID" in
  mike)   API_KEY="T259810e5c9917386b-52c2aeea7cdb5ff41c6685a73cebbeb3" ;;
  howard) API_KEY="Tde5174a6e9e312d14-02fd5bfe0f0ee40c87d027507c680e18" ;;
@@ -122,8 +152,9 @@ fi
 ```bash
 # Write prompt to a workspace path both the Write tool and Git Bash agree on
 # (do NOT use /tmp on Windows — see Hard Rules: /tmp resolves differently in
-# Write vs Git Bash). Use $CLAUDETOOLS_ROOT/.claude/tmp/ or pipe via heredoc.
-PROMPT_FILE="$CLAUDETOOLS_ROOT/.claude/tmp/ollama_prompt.txt"
+# Write vs Git Bash). Use $REPO_ROOT/.claude/tmp/ or pipe via heredoc.
+# $REPO_ROOT is set in the "Get API key" section above.
+PROMPT_FILE="$REPO_ROOT/.claude/tmp/ollama_prompt.txt"
 mkdir -p "$(dirname "$PROMPT_FILE")"
 cat > "$PROMPT_FILE" <<'ENDPROMPT'
 <prompt content here>
@@ -226,6 +257,43 @@ If `OLLAMA` is empty (neither endpoint reachable): Claude drafts the comment bod

 ---

+### Parsing Syncro API Responses
+
+**Problem:** Syncro API responses contain unescaped control characters (U+0000 through U+001F) that break both `jq` and Python's `json.load()`. These characters appear in customer names, ticket descriptions, and other text fields.
+
+**Symptoms:**
+- `jq: parse error: Invalid string: control characters from U+0000 through U+001F must be escaped`
+- Python: `json.decoder.JSONDecodeError: Invalid control character`
+
+**Solution:** Use `grep` and `sed` for simple field extraction instead of jq. For complex parsing where jq is unavoidable, preprocess with `tr -d '\000-\037'` to strip control characters (lossy but functional).
+
+**Common Patterns:**
+
+```bash
+# Extract numeric ID from response (safe - works with control chars)
+TICKET_ID=$(echo "$RESP" | grep -o '"id":[0-9]*' | head -1 | grep -o '[0-9]*')
+TICKET_NUMBER=$(echo "$RESP" | grep -o '"number":[0-9]*' | head -1 | grep -o '[0-9]*')
+CUSTOMER_ID=$(echo "$RESP" | grep -o '"customer_id":[0-9]*' | head -1 | grep -o '[0-9]*')
+
+# Extract string field (use sed to extract between quotes)
+# Example: "status":"New" → New
+STATUS=$(echo "$RESP" | sed -n 's/.*"status":"\([^"]*\)".*/\1/p')
+
+# Extract decimal/float field
+TOTAL=$(echo "$RESP" | grep -o '"total":"[0-9.]*"' | head -1 | grep -o '[0-9.]*')
+
+# Fallback for complex parsing - strip control chars then use jq
+CLEAN=$(echo "$RESP" | tr -d '\000-\037')
+TICKET_ID=$(echo "$CLEAN" | jq -r '.ticket.id')
+```
+
+**When to use each approach:**
+- **grep/sed** (preferred): Extracting numeric IDs, simple string fields, totals
+- **jq with preprocessing** (only when necessary): Complex nested structures, arrays, multiple fields at once
+- **Never retry jq on parse failure** — if jq fails once on a response, it will fail every time with the same input. Switch to grep/sed immediately.
+
+---
+
 ### Adding a per-user key

 1. User logs into Syncro → Admin → API Tokens → New (`/api_tokens/new`)
@@ -946,7 +1014,7 @@ curl -s -X PUT "${BASE}/tickets/${ID}?api_key=${API_KEY}" \
 JSON

 # 5. Bot alert
-bash "$CLAUDETOOLS_ROOT/.claude/scripts/post-bot-alert.sh" \
+bash "$REPO_ROOT/.claude/scripts/post-bot-alert.sh" \
  "[SYNCRO] Mike billed #<number> (<customer>) — ${QTY}h <labor_type>, \$${INVOICE_TOTAL} → https://computerguru.syncromsp.com/tickets/${ID}"
 ```

@@ -974,7 +1042,7 @@ Post after every successful Syncro write. Never post before the write completes.
 **Invocation:**

 ```bash
-ALERT_OUT=$(bash "$CLAUDETOOLS_ROOT/.claude/scripts/post-bot-alert.sh" "<message>")
+ALERT_OUT=$(bash "$REPO_ROOT/.claude/scripts/post-bot-alert.sh" "<message>")
 echo "$ALERT_OUT"
 ```

@@ -1004,20 +1072,20 @@ echo "$ALERT_OUT"

 ```bash
 # Ticket created
-bash "$CLAUDETOOLS_ROOT/.claude/scripts/post-bot-alert.sh" \
+bash "$REPO_ROOT/.claude/scripts/post-bot-alert.sh" \
  "[SYNCRO] Howard created #32301 (Desert Auto Tech) - Server won't boot -> https://computerguru.syncromsp.com/tickets/110736645"
 # Success output: [OK] post-bot-alert: posted to #bot-alerts (message_id=1507055781780918404)

 # Billed + invoiced
-bash "$CLAUDETOOLS_ROOT/.claude/scripts/post-bot-alert.sh" \
+bash "$REPO_ROOT/.claude/scripts/post-bot-alert.sh" \
  "[SYNCRO] Mike billed #32164 (Jerry Burger) - 1.0h remote, \$150.00 -> https://computerguru.syncromsp.com/tickets/110169036"

 # Prepaid billing
-bash "$CLAUDETOOLS_ROOT/.claude/scripts/post-bot-alert.sh" \
+bash "$REPO_ROOT/.claude/scripts/post-bot-alert.sh" \
  "[SYNCRO] Mike billed #32203 (Desert Auto Tech) - 1.5h onsite, applied 1.5 prepay hrs, \$0.00 -> https://computerguru.syncromsp.com/tickets/109895882"

 # Ticket status updated
-bash "$CLAUDETOOLS_ROOT/.claude/scripts/post-bot-alert.sh" \
+bash "$REPO_ROOT/.claude/scripts/post-bot-alert.sh" \
  "[SYNCRO] Mike resolved #32271 (Peaceful Spirit Massage) - IKEv2 VPN setup complete -> https://computerguru.syncromsp.com/tickets/110169036"
 ```

--- a/.claude/scripts/sync.sh
+++ b/.claude/scripts/sync.sh
@@ -54,25 +54,41 @@ TIMESTAMP=$(date "+%Y-%m-%d %H:%M:%S")
 echo -e "${GREEN}[OK]${NC} Starting ClaudeTools sync from $MACHINE at $TIMESTAMP"

 # Navigate to ClaudeTools directory
-# First check: are we already in the repo (or a subdirectory of it)?
-REPO_ROOT=$(git rev-parse --show-toplevel 2>/dev/null || true)
-if [ -n "$REPO_ROOT" ]; then
-    cd "$REPO_ROOT"
-else
-    # Fall back to known candidate paths
+# Read from identity.json (machine-specific, set during onboarding)
+IDENTITY_PATH=""
+for candidate in "$HOME/.claude/identity.json" ".claude/identity.json"; do
+    if [ -f "$candidate" ]; then
+        IDENTITY_PATH="$candidate"
+        break
+    fi
+done
+
+if [ -n "$IDENTITY_PATH" ] && command -v jq >/dev/null 2>&1; then
+    REPO_ROOT=$(jq -r '.claudetools_root // empty' "$IDENTITY_PATH" 2>/dev/null)
+fi
+
+# Fallback: git detection if identity.json doesn't have claudetools_root yet
+if [ -z "$REPO_ROOT" ]; then
+    REPO_ROOT=$(git rev-parse --show-toplevel 2>/dev/null || true)
+fi
+
+# Last resort: hardcoded paths (legacy machines)
+if [ -z "$REPO_ROOT" ]; then
    for candidate in "$HOME/ClaudeTools" "/d/ClaudeTools" "D:/ClaudeTools" "/d/claudetools" "D:/claudetools" "C:/claudetools" "/c/claudetools"; do
        if [ -d "$candidate/.git" ]; then
-            cd "$candidate"
+            REPO_ROOT="$candidate"
            break
        fi
    done
 fi

-if [ ! -d ".git" ]; then
-    echo -e "${RED}[ERROR]${NC} Not in a git working tree"
+if [ -z "$REPO_ROOT" ] || [ ! -d "$REPO_ROOT/.git" ]; then
+    echo -e "${RED}[ERROR]${NC} Cannot locate ClaudeTools repo. Add 'claudetools_root' to identity.json"
    exit 1
 fi

+cd "$REPO_ROOT"
+
 echo -e "${GREEN}[OK]${NC} Working directory: $(pwd)"

 # Detect Python interpreter — verify it actually runs (Windows Store stub passes command -v but fails to execute)