Claude/plan v3 deployment vp mpw - #136
Merged
Merged
Conversation
…urity, 23 Block 1 fixes Menu lifecycle overhaul: - replyMenu() TTL auto-vanish via ttlMs parameter - clearStaleMenuIds() on startup (24h threshold) - mainMenuSentAt / adminMenuSentAt timestamps on user schema - sendHelpMenu() fixed: bare ctx.reply() → replyMenu() (panel stacking bug) - replaceCallbackPanel() fallback now tracks message ID in user state - Admin mode toggle double-fire fixed Feature upgrades: - User giveaway list: sendUserGiveawaysPage() 5/page + 2-min auto-vanish - Admin giveaway panel: activeGiveawaysKeyboard(page) 5/page pagination - pmenu_referral sub-menu with share deep-link button - Admin stats: active window indicator + Refresh button - Admin health panel: inline memory/errors/users/persist-age/uptime - Broadcast builder: Preview button sends to admin DM before mass send Block 1 security & logic fixes: - getStartAppLink(): encodeURIComponent + regex whitelist - referralShareHTML(): escapeHtml() on code param - unwrapTelegramUrl(): safe fallback, scheme whitelist, www.t.me support - getDiscordLink(): null when not configured (no hardcoded fallback) - evaluatePendingActionTimeout(): >= boundary, NaN guard, no createdAt mutation - getPlayLink(): legacy user.playMode fallback restored - Weighted winner pool: splice-after-pick guarantees termination - deploy_status + logs: exec() → execFile() (no shell spawn) - SSHV: rejects null bytes, backticks, $( and ${ before exec - Startup env warnings: ADMIN_IDS, TELEGRAM_CHANNEL_ID, TELEGRAM_GROUP_ID Centralized helpers: computeParticipantWeight(), getRealGiveaways(), isNewUserPromoEligible() Metrics: added runewager_menu_stale_recoveries, pending_actions_timed_out, uptime_seconds load_tooltips.sh: full rewrite — atomic write, --push/--pull flags, parameterized REPO_DIR, HTML-safe tooltips, --dry-run mode, permission checks, guarded git ops Tests: readdirSync wrapped in try/catch; isCatchAllRegexPattern expanded; extractCommandHandlerNames supports let/var/no-semicolon Infrastructure: pre-deploy-checks gate 3b (menu symbols), CHANGELOG.md created, package.json bumped to 3.0.0 All 60 tests pass. node --check clean. https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
… during PR merges Restores 12 categories of codex changes that were incorrectly overwritten by "current" (main) when resolving merge conflicts across 6 merge commits. Security / TLS hardening (index.js): - Restore resolveTlsCertPathIfAllowed(): multi-directory fallback for TLS certs (PROJECT_DIR, certs/, /etc/ssl, /etc/letsencrypt) — was collapsed to PROJECT_DIR only - Restore isHealthTlsEnabled() to use resolveTlsCertPathIfAllowed() - Restore requestHealthPayload(): centralized HTTP/HTTPS health fetcher with dynamic protocol detection — was inlined with hardcoded https in two places - Update /health command to use requestHealthPayload() (shows protocol label) - Update pamenu_tools_health to use requestHealthPayload() - Update startHealthServer() TLS cert read to use resolveTlsCertPathIfAllowed() Command injection protection (index.js): - commandNeedsConfirmation(): restore pipe-to-bash/zsh detection, destructive-cmd-piped pattern, redirect pattern — main simplified to only catch "sh" - commandBlocked(): restore explicit pipe-to-shell blocker pattern Race condition fixes (index.js): - deleteEphemeralBonusPrompt(): restore runUserMutation guards for safe read-delete-write of ephemeralBonusMsgId/ChatId - sendEphemeralBonusPrompt(): restore guards (claimedPromo check, active promo lookup), per-user dynamic promo lookup via getActivePromoCodeForUser() (was hardcoded promoStore.code), "I Have Claimed" callback button, mutation-safe writes - Add promo_confirm_claimed_next callback handler (button was added, handler missing) Deploy/runtime hardening (deploy.sh, prod-run.sh, runewager.service): - deploy.sh: restore data/backups/ mkdir, chown -R APP_NAME, chmod 0750/0640/0600 permission hardening after npm install - prod-run.sh: restore chmod 0750 for data/data/backups/logs dirs, chmod 0640 for log+session files, chown -R to service user after dir creation - runewager.service: add UMask=0077 (prevents world-readable files from service) .gitignore: - Restore legacy root-level data file patterns: users.json, giveaways.json, promo.json, env.json, analytics*.json (pre-data/ directory structure) All 60 tests pass. node --check clean. https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
- TDZ fix: move pendingActionsTimedOut/menuStaleRecoveries declarations
before evaluatePendingActionTimeout to eliminate temporal dead-zone risk
- Timeout boundary: revert < to <= so age exactly equal to 15m is NOT
expired — aligns with /testall check and unit test expectations
- /logs line count: clamp to [1, 200] (adds Math.max(1,...) lower bound
to reject negative/zero values from user input)
- Health panel: fix _errorRate.windowErrors → _errorRate.count (field
did not exist; .count is the correct field on _errorRate object)
- /logs fallback: add execFile('tail') fallback when journalctl errors
with no output (non-systemd systems); uses BOT_LOG_FILE env or default
- executeSshvCommand: fix comment — said "spawn" but code uses exec();
updated to accurately describe exec with shell:true (admin-only)
- load_tooltips.sh: remove || true that defeated diff --cached check,
causing .gitignore changes to never be committed on --push
All 60 tests pass. node --check clean.
https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
## Helpful Tooltips (was Content Drops) - Rename all UI strings: Content Drops → Helpful Tooltips throughout - New settings panel: interval, silent mode, Link Channel/Group - Target group linking via forwarded message (saves name + ID) - Dashboard footer shows real group name + ID - "Show all Helpful Tooltips (N)" button with dynamic count - Inline button builder: [Label - URL] && [Label2 - URL2] syntax - Multiple buttons per row with &&, new line = new row - [Open Bot] shorthand for standard Open Bot button - Full URL + label validation before save - postTipToConfiguredTarget uses silentMode flag + parsed buttons - tipsStore extended with targetGroupTitle and silentMode fields ## Giveaway v3.0+ - Extracted buildGiveawayAnnouncementText + buildGiveawayAnnouncementKeyboard helpers - scheduleGiveawayRefresh: auto-refresh at 25%, 50%, 75% of duration + re-pin - scheduleGiveawayReminders overhauled: 10m, 5m, 1min, 30sec, 10→1 countdown - HTML results format: @handle, SC WON, (2x boost applied), DM tip - Full admin winner report per winner: TG handle, display name, RW username, prize, boost - "View Results in Group" deep-link button in admin report - Winner DMs include SC amount, boost status, RW username - DM failure tracking with summary count - giveawayPreflightCheck: validates group linked, warns on missing pin permission - gwizStart calls preflight before wizard begins ## Scripts - generate_tooltips.sh: extracts DEFAULT_TIPS_LIST from index.js → data/tooltips.json (atomic, idempotent) - add_tooltip.sh: appends placeholder tooltip entry, outputs new ID - deploy.sh: step 3b auto-runs generate_tooltips.sh before service start - prod-run.sh: step 6b auto-runs generate_tooltips.sh before bot launch ## /testall - Added Helpful Tooltips System checks (tipsStore shape, count, interval, target, parser) - Added Giveaway Extended checks (helpers defined, preflight defined) ## Docs - RUNEWAGER_FUNCTIONALITY_MAP.md: full v3.0+ sync with flowcharts All 60 tests pass. node --check clean. https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
… linking - Group command guard middleware: bot.use() intercepts all commands in group/supergroup chats and redirects to DM with a deep-link button. Passthrough commands with own group logic: link, linkrunewager, giveaway, start_giveaway, admin. Suppresses handler execution for all others. - Onboarding progress bar: onboardingProgressBar(step) renders ●●○○○ Step N of 5 — Label. showOnboardingPrompt() prepends a Markdown progress header (auto-deletes after 8s) before each step-specific prompt. - Onboarding completion card: shown once (user.onboarding.completionCardShown flag) when user reaches the main menu after completing all 5 steps. Includes feature summary and Open Menu button. - Admin System Tools: added 🔗 Group Linking button to adminSystemToolsKeyboard() with admin_sys_group_linking action handler (renders group linking panel with back-to-system-tools navigation). - Schema: completionCardShown added to onboarding default + migration guard. - Map: RUNEWAGER_FUNCTIONALITY_MAP.md fully updated; todolist.md updated. - All 60 tests pass, node --check clean. https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
…ates PR #112 review fixes (sourcery-ai): - R1: tips_cmd_import_batch now uses await_tip_import_batch pending type with dedicated JSON-array router handler; proper MarkdownV2 prompt - R2: generate_tooltips.sh fixes command substitution pollution; use RUNEWAGER_APP env var instead of process.argv[1] (undefined in node -) - R3: add_tooltip.sh fixes shell injection; TOOLTIP_TEXT passed via TOOLTIP_TEXT_ENV env var, heredoc uses <<'EOF', process.argv[2] for file - R4: extend catchAllCases array with (.|\n)*, (.|\n)+, (\.|[\s\S])*; add post-whitespace-strip forms to CATCH_ALL_CORES set - R5: extractCommandHandlerNames test now covers let/var declarations and no-semicolon forms (CMD_FOUR/eta, CMD_FIVE/theta) - R6: RUNEWAGER_FUNCTIONALITY_MAP.md typo "auto-deletes 8s" → "after 8s" Codebase audit duplicate removal: - A1: remove dead buildGiveawayAnnouncementText(giveaway, remainingStr) at ~12533; keep dynamic version at ~13795 - A2: remove simplified buildGiveawayAnnouncementKeyboard at ~13817 (wrong tgw_participants_ callback); restore full 5-row version - A3+A4: remove first duplicate bot.action registrations for admin_cat_system and admin_cat_support (identical bodies) All 60 tests pass, node --check clean, bash -n clean on both scripts. https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
Resolved conflicts in favour of our branch (HEAD) for: - index.js: keep await_tip_import_batch router + handler, keep A1/A2 duplicate-function removal, keep Content Drops branding from main - add_tooltip.sh / generate_tooltips.sh: keep env-var injection fixes - test/smoke.test.js: keep extended catchAllCases + extractCommand fixtures - RUNEWAGER_FUNCTIONALITY_MAP.md / todolist.md: keep our updated content Incoming from origin/main: claude-pr-results-20260228_193756Z.md https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
…ry bot start Every script that starts/restarts the bot now follows the same safe sequence: 1. git pull origin main (fetch + reset --hard) 2. generate_tooltips.sh (extracts DEFAULT_TIPS_LIST → data/tooltips.json) 3. kill any process blocking PORT (default 3000) 4. start/restart bot Changes per file: - prod-run.sh: add step 9c — free_port_if_conflicted() BEFORE step 10 restart (port-kill was already present in God-Mode Heal but fired after, not before) - deploy.sh: add step 3c — inline lsof/fuser port-kill before systemctl start - start.sh: add git fetch+reset, generate_tooltips, port-kill, stale-PID kill before bot launch; replace refuse-on-duplicate with kill-and-continue - dev-run.sh: add git fetch+reset (best-effort), generate_tooltips, port-kill before exec node - scripts/rollback.sh: add generate_tooltips after npm ci (refreshes from rolled-back index.js), add lsof/fuser port-kill before service start (no git pull — rollback intentionally targets an older commit) https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
- Escape Markdown special chars in runewagerUsername to prevent parse failures - Add explicit smoke-test assertions for (.|)* and (.|)+ catch-all forms - Strip inline comments from .env PORT values in deploy.sh, dev-run.sh, start.sh (e.g. PORT=3000 # dev now correctly yields 3000) - Fix showOnboardingPrompt JSDoc: steps documented as 1–4 → 1–5 (matches impl) - Upgrade all port-block kill loops to SIGTERM-first then SIGKILL after 2s grace - Add path + shape validation to generate_tooltips.sh Node extraction block (validates RUNEWAGER_APP is absolute .js path; checks array literal size/shape) https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
- Create docs/INDEX.md: exhaustive cross-reference of all 95 commands, 266 action handlers, and 50+ pending action types mapped to feature docs - Create docs/features/ with 15 per-feature .md files covering every bot subsystem (onboarding, menus, giveaway, bonus, promos, tooltips, referral, SSHV, deploy, user lookup, group linking, bug reports, announcements, misc) - Create docs/TODO_FUNCTIONALITY_UPGRADE.md with 14 open stale-menu and missing-handler items (walkthrough dead-end, clearOldMenus gaps, missing tip_view handler, language stub, broadcastFailedUsers cap) - Update RUNEWAGER_FUNCTIONALITY_MAP.md section 26 with full docs/ table and mandate for future Claude sessions to consult docs/INDEX.md first Future sessions: read docs/INDEX.md → feature .md → index.js (if needed) https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
… path validation - index.js: keep MarkdownV2 escaping for runewagerUsername (security fix) - generate_tooltips.sh: keep path validation (absolute .js, no traversal) - start.sh / dev-run.sh / scripts/rollback.sh: keep SIGTERM→SIGKILL two-step - start.sh: keep inline comment stripping in PORT parse - test/smoke.test.js: keep (.|)* and (.|)+ catch-all cases https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
PR review comment fixes: - add_tooltip.sh: validate list is Array, filter non-finite IDs before computing maxId; throw fast on malformed tooltips.json - docs/12-group-linking.md: fix entry points table — admin System Tools uses admin_sys_group_linking, not admin_cat_system - docs/15-misc-commands.md: remove⚠️ walkthrough dead-end note (resolved) T-01 — Walkthrough system hardened: - sendWalkthroughStep(): clearOldMenus() before every step send - Back button disabled on step 1 (first step) - Next replaced with Finish button on last step (step 35) - walk_done on last step: sets started=false, fires walkthrough_completed analytics, answerCbQuery with success toast, returns to main menu - New doc: docs/features/16-walkthrough.md T-02 — clearOldMenus() added to 5 missing locations: - sendOnboardingReferralPrompt() - renderSshvConsole() - renderGroupLinkingTools() - tips_cmd_edit handler - tips_cmd_remove handler (sendGiveawayListPage already used replyMenu() — no change needed) T-03 — Tooltip view flow implemented: - tipsDashboardKeyboard(): "👁 View Tooltip" button added (row 4) - tips_cmd_view handler: clearOldMenus + tipSelectKeyboard('tip_view') - tip_view_{id} handler: preview card with Prev/Next tip navigation, Edit / Toggle / Delete buttons, Back to List, Admin Menu T-15 — Broadcast failure logging made reliable: - Removed 500-item cap from broadcastFailedUsers push and load - Every failure logged via adminLog('broadcast_failure', ...) to data/admin-events.log — no silent drops - /broadcast_failed: chunked output (30/chunk, up to 90 inline); overflow note points to log file - High-failure-rate warning: >20% failure rate sends⚠️ DM to ADMIN_IDS Merge conflicts resolved (PRs #112-114): - index.js: keep MarkdownV2 escaping for runewagerUsername - generate_tooltips.sh: keep path validation (absolute .js, no traversal) - start.sh/dev-run.sh/rollback.sh: keep SIGTERM→SIGKILL two-step - test/smoke.test.js: keep (.|)* and (.|)+ catch-all test cases Tests: 60/60 pass | node --check: clean | bash -n: all scripts OK https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
Three changes addressing remaining PR #115 review feedback: 1. generate_tooltips.sh — skip guard to preserve runtime tips - Skip regeneration if data/tooltips.json already has entries. - Prevents admin-added tooltips (broadcast every 4h via bot) from being overwritten on every restart or deploy. - Add --force flag to override the guard when intentional reset is needed. - --dry-run still works regardless of the guard. 2. scripts/helpers/free_port.sh — shared SIGTERM→SIGKILL helper - Extracts duplicated port-freeing logic (lsof/fuser, SIGTERM→SIGKILL) from start.sh, dev-run.sh, deploy.sh, and scripts/rollback.sh into a single reusable helper (sourceable or callable directly). - Reduces drift: future tweaks to kill strategy happen in one place. 3. start.sh / dev-run.sh — RUNEWAGER_AUTO_UPDATE gate - git fetch + reset --hard origin/main is now conditional on RUNEWAGER_AUTO_UPDATE env var (default 1 in prod start.sh, default 0 in dev-run.sh). - Prevents silent discard of local/staging uncommitted changes. - Documented in .env.example. https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
1. generate_tooltips.sh — normalize APP_DIR to absolute path Move helper function definitions (info/warn/error) before variable assignments so error() is available at init time. Normalize RUNEWAGER_DIR → APP_DIR via cd+pwd immediately after assignment so the Node.js absolute-path validation (requires '/'-prefixed path) never fails when a caller passes a relative RUNEWAGER_DIR. 2. scripts/helpers/free_port.sh — re-query port before SIGKILL Extract discovery into _query_port_pids() helper. After the SIGTERM grace period, re-query the port for survivors and only SIGKILL PIDs that are still listening — guards against killing an unrelated process that reused a PID during the 2 s sleep window. 3. dev-run.sh — read RUNEWAGER_AUTO_UPDATE from .env as fallback Parse RUNEWAGER_AUTO_UPDATE from .env before the auto-update guard so the flag works even when .env values have not been exported into the calling shell. Use explicit if/else instead of chained && || for the destructive git reset --hard command. 4. start.sh — same .env-read fix + explicit if/else for git reset Same pattern as dev-run.sh: resolve RUNEWAGER_AUTO_UPDATE from env then .env (default 1 for prod), replace the chained git &&/|| with an explicit if/else block. https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
1. generate_tooltips.sh — case-based arg parser Replace the for-loop [[]] checks with a case statement that rejects unknown flags (e.g. --froce typo) with a clear error and non-zero exit. 2. generate_tooltips.sh — env-var file paths in Node.js invocations The three inline node -e calls that interpolated $TOOLTIPS_FILE / $TMP_FILE directly into single-quoted JS strings were fragile for paths containing quotes or special characters. All three now pass the path via a dedicated env var (TOOLTIPS_FILE_PATH or VALIDATE_FILE) and read process.env inside the script, matching the existing RUNEWAGER_APP pattern. 3. dev-run.sh — non-fatal free_port.sh invocation free_port.sh can exit non-zero on benign errors (no lsof/fuser, race after SIGTERM) which would abort dev-run.sh under set -eu. Added || echo WARN fallback to mirror the same non-fatal pattern used for the tooltip script invocation directly above. https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
1. generate_tooltips.sh — trap for TMP_FILE cleanup Register 'trap rm -f TMP_FILE EXIT INT TERM' immediately before the atomic write section so the temp file is always removed on any exit (error, signal, or normal completion). After a successful mv the path no longer exists, so the trap is a safe no-op on the happy path. 2. dev-run.sh — default to merge --ff-only; gate reset --hard behind opt-in Auto-update now runs 'git fetch + merge --ff-only' (non-destructive). 'git reset --hard origin/main' is only executed when RUNEWAGER_FORCE_RESET=1 is set in the environment or .env, satisfying the "confirm destructive operations" guideline. Fast-forward failure emits a clear warning pointing the user to RUNEWAGER_FORCE_RESET. Documented in .env.example. https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
- Merged main branch: telegramSafe.js, rateLimiter.js, backend.js, runewager-endpoint.service, prod-run.sh rewrite, runewager_redeploy.sh, rw_cpu_guard.sh - Removed dead code: legacy adminKeyboard() function (JSDoc + body, ~32 lines) — no callers, belonged to removed /admin_menu command - RUNEWAGER_FUNCTIONALITY_MAP.md: updated last-audited date, added new module entries (telegramSafe, rateLimiter, backend, service, scripts), added 2026-03-04 audit log entry - todolist.md: updated last-updated date, added fixed adminKeyboard entry - All 60 tests pass post-fix https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
…mode Three code-review fixes: 1. Section 10 (Safe restart): `|| true` swallowed systemctl failures and left the bot stopped. Replaced with `if ! systemctl restart ...; then` block that falls back to manual kill + nohup when systemd fails. 2. Bare `disown` (non-systemd path, L506): with `set -euo pipefail` a failed `disown` (no job control in non-interactive shells) aborted the script before post-start health checks and Telegram reporting ran. Fixed: `disown || true` in both the fallback and non-systemd paths. 3. Telegram notification: removed `parse_mode=Markdown` (unescaped log content and env values can break Markdown rendering / cause truncation). Switched to plain text with `--data-urlencode` so special chars in the message are safe without manual escaping. Removed unused `_REPORT` variable (log tail was computed but never injected into the message). 60/60 tests pass. https://claude.ai/code/session_01VijmtzjN63WZJy5gYgJAKs
|
CodeAnt AI is reviewing your PR. Thanks for using CodeAnt! 🎉We're free for open-source projects. if you're enjoying it, help us grow by sharing. Share on X · |
Reviewer's GuideCleans up a dead admin promo keyboard helper, hardens the production restart script with a systemd failure fallback and safer disown usage, and updates audit/documentation metadata to reflect the latest /runewager-audit pass and new backend/runtime components. Sequence diagram for deploy restart and Telegram notificationsequenceDiagram
actor Admin
participant prod_run_sh
participant systemd
participant NodeProcess
participant TelegramAPI
Admin->>prod_run_sh: Execute prod-run.sh
prod_run_sh->>prod_run_sh: Detect existing PID
alt systemd available and service file exists
prod_run_sh->>systemd: restart APP_NAME.service
alt systemd restart succeeds
systemd-->>prod_run_sh: Service restarted
else systemd restart fails
prod_run_sh->>prod_run_sh: Log warning and fallback
prod_run_sh->>NodeProcess: kill existing PID (if any)
prod_run_sh->>NodeProcess: start via nohup node index.js
prod_run_sh->>prod_run_sh: disown with failure ignored
end
else systemd not available or service file missing
prod_run_sh->>NodeProcess: kill existing PID (if any)
prod_run_sh->>NodeProcess: start via nohup node index.js
prod_run_sh->>prod_run_sh: disown with failure ignored
end
prod_run_sh->>prod_run_sh: Sleep and refresh PID, run health checks
prod_run_sh->>prod_run_sh: Build plain text _MSG without parse_mode
loop For each admin id
prod_run_sh->>TelegramAPI: sendMessage(chat_id, text) via URL-encoded POST
TelegramAPI-->>prod_run_sh: HTTP response (ignored on failure)
end
Class diagram for removal of legacy adminKeyboard helperclassDiagram
class index_js {
<<module>>
+linkPrefKeyboard()
%% adminKeyboard() removed as dead code
}
Flow diagram for updated prod-run.sh restart logicflowchart TD
A[Start prod-run restart section] --> B{Existing PID found}
B -- yes --> C[Log: Safe restart of running bot]
B -- no --> D{systemctl available and service file exists}
C --> D
D -- yes --> E{systemctl restart APP_NAME.service succeeds}
D -- no --> H[Kill existing PID if present]
E -- yes --> I[Continue after restart]
E -- no --> F[Log warning: systemctl restart failed]
F --> G[Kill existing PID if present]
G --> J[Sleep 1s]
J --> K[Start bot via nohup node index.js]
K --> L[disown with failure ignored]
L --> I
D -- no --> H
H --> M[Start bot via nohup node index.js]
M --> N[disown with failure ignored]
N --> I
I --> O[Sleep 3s and refresh PID]
O --> P[Continue to health checks and notifications]
File-Level Changes
Tips and commandsInteracting with Sourcery
Customizing Your ExperienceAccess your dashboard to:
Getting Help
|
|
Caution Review failedThe pull request is closed. ℹ️ Recent review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (4)
📝 WalkthroughWalkthroughThis PR documents v3.1 audit completion (0 critical issues, 1 warning resolved), removes the legacy adminKeyboard() function from index.js, hardens the production restart script with fallback logic for failed systemd restarts, and converts admin notifications from Markdown to plain-text URL-encoded format. Changes
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Possibly related PRs
Suggested labels
Poem
✨ Finishing Touches
🧪 Generate unit tests (beta)
Tip Try Coding Plans. Let us write the prompt for your AI agent so you can ship faster (with fewer bugs). Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
codeant-ai
Bot
added
the
size:S
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
Nitpicks 🔍
|
![codeant-ai[bot]](https://avatars.githubusercontent.com/in/646884?s=60&v=4){kind=small}
| # Telegram admin notification (plain text — no parse_mode to avoid Markdown rendering issues | ||
| # with unescaped log content or special chars in env values) | ||
| _ADMIN_IDS="$(read_env_value ADMIN_IDS || true)" | ||
| _BOT_TOKEN="$(read_env_value TELEGRAM_BOT_TOKEN || true)" |
There was a problem hiding this comment.
Suggestion: The deploy-complete Telegram notification only reads the legacy TELEGRAM_BOT_TOKEN from .env and ignores the canonical BOT_TOKEN (and any exported BOT_TOKEN env var), so on setups that follow the documented convention of using BOT_TOKEN only, the health summary message will never be sent even though the bot runs correctly. [logic error]
Severity Level: Major ⚠️
- ⚠️ prod-run.sh post-deploy Telegram summary skipped with BOT_TOKEN-only.
- ⚠️ Ops lose automatic health summary after each deployment.
Suggested change
| _BOT_TOKEN="$(read_env_value TELEGRAM_BOT_TOKEN || true)" | |
| # Prefer canonical BOT_TOKEN (env or .env), fall back to legacy TELEGRAM_BOT_TOKEN alias | |
| _BOT_TOKEN="${BOT_TOKEN:-$(read_env_value BOT_TOKEN || read_env_value TELEGRAM_BOT_TOKEN || true)}" |
Steps of Reproduction ✅
1. Configure the bot using the documented canonical token: in `/workspace/Runewager/.env`,
set `BOT_TOKEN=...` and `ADMIN_IDS=...` while leaving `TELEGRAM_BOT_TOKEN` unset or empty
(README recommends `BOT_TOKEN` at `README.md:15` and `.env.example` declares `BOT_TOKEN`
as the canonical name at `CLAUDE_FIX_PR.md:2474`).
2. Start the bot normally (e.g., via systemd using the unit generated by `prod-run.sh`)
and confirm it runs: `index.js:14` defines `const BOT_TOKEN = process.env.BOT_TOKEN ||
process.env.TELEGRAM_BOT_TOKEN;`, so with only `BOT_TOKEN` set the application starts
successfully and serves the health endpoint.
3. Run the deployment runner `prod-run.sh` from `/workspace/Runewager` as intended for
production deployments; it performs the restart and post-start health check (section 10
and 11, `prod-run.sh:496-553`), and the health check succeeds so the script reaches the
Telegram notification block at `prod-run.sh:556-569`.
4. In that notification block, `_ADMIN_IDS` is populated from `.env`, but `_BOT_TOKEN` is
set only via `read_env_value TELEGRAM_BOT_TOKEN` (`prod-run.sh:558-559`), which returns
empty when `TELEGRAM_BOT_TOKEN` is not configured; the condition `[[ -n "$_BOT_TOKEN" &&
-n "$_ADMIN_IDS" ]]` at `prod-run.sh:560` fails, so the `curl` loop at
`prod-run.sh:563-567` never executes and no "Deploy complete" Telegram messages are sent,
despite the bot and health checks working correctly with `BOT_TOKEN`.Prompt for AI Agent 🤖
This is a comment left during a code review.
**Path:** prod-run.sh
**Line:** 559:559
**Comment:**
*Logic Error: The deploy-complete Telegram notification only reads the legacy `TELEGRAM_BOT_TOKEN` from `.env` and ignores the canonical `BOT_TOKEN` (and any exported `BOT_TOKEN` env var), so on setups that follow the documented convention of using `BOT_TOKEN` only, the health summary message will never be sent even though the bot runs correctly.
Validate the correctness of the flagged issue. If correct, How can I resolve this? If you propose a fix, implement it and please make it concise.
|
CodeAnt AI finished reviewing your PR. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
User description
Summary by Sourcery
Remove unused admin promo keyboard and harden production restart and deploy notification behavior.
Bug Fixes:
adminKeyboard()promo keyboard function that no longer has any callers.Enhancements:
prod-run.shrestart logic to fall back to manual kill and nohup start whensystemctlrestart fails.disownusage inprod-run.shto avoid errors when nohup backgrounding fails.Documentation:
adminKeyboard()as a resolved bug.CodeAnt-AI Description
Remove legacy admin promo keyboard and make deploy/restart behavior more robust and observable
What Changed
Impact
✅ Fewer failed restarts during deployments✅ Clearer deploy notifications to admins (no Markdown rendering errors)✅ Reduced legacy dead-code reported in audits💡 Usage Guide
Checking Your Pull Request
Every time you make a pull request, our system automatically looks through it. We check for security issues, mistakes in how you're setting up your infrastructure, and common code problems. We do this to make sure your changes are solid and won't cause any trouble later.
Talking to CodeAnt AI
Got a question or need a hand with something in your pull request? You can easily get in touch with CodeAnt AI right here. Just type the following in a comment on your pull request, and replace "Your question here" with whatever you want to ask:
This lets you have a chat with CodeAnt AI about your pull request, making it easier to understand and improve your code.
Example
Preserve Org Learnings with CodeAnt
You can record team preferences so CodeAnt AI applies them in future reviews. Reply directly to the specific CodeAnt AI suggestion (in the same thread) and replace "Your feedback here" with your input:
This helps CodeAnt AI learn and adapt to your team's coding style and standards.
Example
Retrigger review
Ask CodeAnt AI to review the PR again, by typing:
Check Your Repository Health
To analyze the health of your code repository, visit our dashboard at https://app.codeant.ai. This tool helps you identify potential issues and areas for improvement in your codebase, ensuring your repository maintains high standards of code health.
Summary by CodeRabbit
Bug Fixes
Chores