social-autoposter 1.6.7 → 1.6.9

package/bin/cli.js

@@ -165,6 +165,13 @@ function writeAppMakerEnvFile() {
     '# 9222 is the one the user logs into via the AppMaker UI.',
     'export TWITTER_CDP_URL="http://127.0.0.1:9222"',
     '',
+    '# AppMaker VMs run as root and the appmaker template sets Claude defaultMode',
+    '# to bypassPermissions. Claude CLI refuses bypassPermissions under root for',
+    '# security reasons UNLESS IS_SANDBOX=1 is set. Without this, every `claude -p`',
+    '# call in the pipeline exits immediately with no output (cost=$0.00, 16s) and',
+    '# Phase 1 reports envelope parse error / phase1_no_tweets.',
+    'export IS_SANDBOX=1',
+    '',
   ].join('\n');
   fs.writeFileSync(envPath, body);
   console.log(`  AppMaker VM detected -> wrote ${envPath} (TWITTER_CDP_URL=http://127.0.0.1:9222)`);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "social-autoposter",
-  "version": "1.6.7",
+  "version": "1.6.9",
   "description": "Automated social posting pipeline for Reddit, X/Twitter, LinkedIn, and Moltbook. Install as a Claude Code agent skill.",
   "bin": {
     "social-autoposter": "bin/cli.js"

package/skill/run-twitter-cycle.sh

@@ -634,7 +634,7 @@ $HOOK_NOTICE
 $ANTI_DEBUG_RULE
 
 Query guidelines:
-- MANDATORY: every query MUST end with the operator \`since_time:$(date -u -v-${FRESHNESS_HOURS}H +%s)\` copied EXACTLY as written. It is a pre-computed Unix-epoch timestamp; do NOT recalculate, reformat, or round it, just paste it verbatim into every single query. It restricts X to tweets posted in the last ${FRESHNESS_HOURS} hours, which is the cycle's freshness wall: tweets older than that are dropped at scoring and the whole search is wasted. Do NOT use the day-granular \`since:YYYY-MM-DD\` operator: it admits tweets up to ~45h old that the scorer then discards. Even if some past top-performing queries shown below still use \`since:\`, you MUST use \`since_time:\` instead; those examples predate this rule.
+- MANDATORY: every query MUST end with the operator \`since_time:$(( $(date +%s) - FRESHNESS_HOURS * 3600 ))\` copied EXACTLY as written. It is a pre-computed Unix-epoch timestamp; do NOT recalculate, reformat, or round it, just paste it verbatim into every single query. It restricts X to tweets posted in the last ${FRESHNESS_HOURS} hours, which is the cycle's freshness wall: tweets older than that are dropped at scoring and the whole search is wasted. Do NOT use the day-granular \`since:YYYY-MM-DD\` operator: it admits tweets up to ~45h old that the scorer then discards. Even if some past top-performing queries shown below still use \`since:\`, you MUST use \`since_time:\` instead; those examples predate this rule.
 - MANDATORY EVEN IF YOUR QUERY KEYWORDS DO NOT NAME THE EXCLUDED TOPIC. If a project's \`excludes_for_search\` array is non-empty, append \`-term\` for EVERY listed term to that project's query, verbatim, with NO EXCEPTIONS. The exclusion is project-wide and persistent — it is a safety rail against ALL future false positives for that project, not a query-keyword-conditional rule. Do NOT reason "my search keywords are about meditation/local AI/vibe coding so the cricket/crypto/Bolt excludes are unnecessary" — that reasoning defeats the rail. The terms passed the >=2-batch activation gate; they have ALREADY survived a one-off filter. Your job here is purely mechanical concatenation, not editorial judgment. Concrete examples of what MUST happen: if Vipassana has \`excludes_for_search: ["cricket","ipl","kohli","lsg","pant","csk","inglis","suvendu","tmc","bjp"]\`, your Vipassana query MUST end with \` -cricket -ipl -kohli -lsg -pant -csk -inglis -suvendu -tmc -bjp\` even when the query searches for "meditation" or "vipassana" with no mention of Goenka. If fazm has \`excludes_for_search: ["memecoin","okx","onchain"]\`, every fazm query gets \` -memecoin -okx -onchain\` appended whether searching "local AI", "RPA", or anything else. If mk0r has \`excludes_for_search: ["usain"]\`, every mk0r query gets \` -usain\`. Skipping these in any query is a bug.
 - MANDATORY: pick \`min_faves:N\` per the PER-PROJECT SUPPLY SIGNAL above. If the project has no entry in the supply table (new project or first cycle), use min_faves:20 as a starting point; the next cycle will see your attempt and self-tune. Never hardcode min_faves:50 for a project whose supply table shows zero results at that tier.
 - Favor discussions/opinions (people sharing experience, asking questions), not news/promos/giveaways