feat(claws): add unified sageox OpenClaw skill#574
Conversation
Single skill with 8 capabilities replacing separate sageox-distill and sageox-summary skills: query, coworkers, distill, distill-pipeline, summary, glance, catchup, and import/export. Tested 10/10 on VPS. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-Authored-By: SageOx <ox@sageox.ai> SageOx-Session: https://sageox.ai/repo/repo_019c5812-01e9-7b7d-b5b1-321c471c9777/sessions/2026-04-28T17-29-avi-OxUnkE/view
|
Warning Rate limit exceeded
To continue reviewing without waiting, purchase usage credits in the billing tab. ⌛ How to resolve this issue?After the wait time has elapsed, a review can be triggered using the We recommend that you space out your commits to avoid hitting the rate limit. 🚦 How do rate limits work?CodeRabbit enforces hourly rate limits for each developer per organization. Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout. Please see our FAQ for further information. ℹ️ Review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (2)
📝 WalkthroughWalkthroughAdds the SageOx OpenClaw toolkit: documentation, templates, installer and readiness scripts for a pinned ChangesSageOx Toolkit (single coherent change DAG)
Sequence Diagram(s)sequenceDiagram
autonumber
participant User as User
participant FS as Filesystem/Repo
participant Ox as ox CLI
participant Claude as Claude
User->>Ox: run "catchup" (time window)
Ox->>FS: ox distill/history, ox session list, ox insights
FS-->>Ox: gathered artifacts (history, sessions, insights)
Ox->>FS: render assets/CATCHUP.md with artifacts
Ox->>Claude: timeout 600 claude -p (prompt with templated content)
Claude-->>Ox: synthesized briefing (mrkdwn)
Ox-->>User: briefing (or error / raw-data offer)
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Tip 💬 Introducing Slack Agent: The best way for teams to turn conversations into code.Slack Agent is built on CodeRabbit's deep understanding of your code, so your team can collaborate across the entire SDLC without losing context.
Built for teams:
One agent for your entire SDLC. Right inside Slack. 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 |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 10
🧹 Nitpick comments (2)
claws/openclaw/sageox/deploy-to-vps.sh (1)
8-8: Using$HOMEis clearer, but the quoted tilde will still expand correctly in rsync/SSH.When rsync sends a remote path via SSH (format
host:path), the remote shell expands tilde syntax. Since the connection is as theubuntuuser,~will expand to/home/ubuntuon the VPS, and deployment will work correctly.Using
$HOMEis still recommended for code clarity and consistency withVPS_KEYabove it, but this is not a functional issue.Suggested improvement
-SKILL_DIR="~/.openclaw/workspace/skills/sageox" +SKILL_DIR="$HOME/.openclaw/workspace/skills/sageox"🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@claws/openclaw/sageox/deploy-to-vps.sh` at line 8, The SKILL_DIR variable in deploy-to-vps.sh uses a quoted tilde ("~/.openclaw/...") which works but is less explicit; change SKILL_DIR to use $HOME (e.g., SKILL_DIR="$HOME/.openclaw/workspace/skills/sageox") for clarity and consistency with VPS_KEY and other env vars, ensuring the path expands correctly in local and remote contexts (refer to the SKILL_DIR assignment in deploy-to-vps.sh).claws/openclaw/sageox/references/setup.md (1)
72-74: ⚡ Quick winAdd concrete verification command for claude credentials.
Steps 1-3 of the authentication checklist provide explicit verification commands (
ox status,gh auth status,git config user.name), but step 4 only describes the credential sources without showing how to verify. Consider adding a concrete verification step for consistency:4. Confirm `claude` has credentials: ```bash claude --version && (test -f ~/.claude/config.json || test -n "$ANTHROPIC_API_KEY")If this fails, either run
claude login(Pro/Max OAuth) or exportANTHROPIC_API_KEYin the shell that launches OpenClaw.This maintains the pattern established by the other checks and gives users a clear pass/fail signal. <details> <summary>🤖 Prompt for AI Agents</summary>Verify each finding against the current code and only fix it if needed.
In
@claws/openclaw/sageox/references/setup.mdaround lines 72 - 74, Add a
concrete verification command for Claude credentials: update the checklist step
mentioningclaude/ANTHROPIC_API_KEYto include a shell check that verifies
theclaudeCLI is available and that either~/.claude/config.jsonexists or
theANTHROPIC_API_KEYenvironment variable is set; if the check fails,
instruct to runclaude loginor exportANTHROPIC_API_KEYbefore launching
OpenClaw.</details> </blockquote></details> </blockquote></details> <details> <summary>🤖 Prompt for all review comments with AI agents</summary>Verify each finding against the current code and only fix it if needed.
Inline comments:
In@claws/openclaw/sageox/README.md:
- Around line 18-22: Update the Prerequisites section to include the missing ox
bootstrap/prime steps: explicitly list and order the commands users must run
(start withox agent primefirst, thenox login,ox init,ox doctor, and
finallyox status) and add a short note thatox agent primemust be run
before any other action to enable SageOx queries; ensure these commands are
placed alongside the existingoxinstallation note so users see installation +
required initialization steps together.In
@claws/openclaw/sageox/references/summary.md:
- Around line 27-29: The fenced code blocks in the summary (the warning line,
the "### Team" block and the bulleted list block — look for the triple-backtick
blocks surrounding the warning string, the "### Team "<team_id>"" section and
the bullets starting with "- Organize the summary...") are missing language
identifiers; update each opening fence fromtotext so they become
text ...which satisfies MD040 and clarifies the block type without
changing content.- Around line 75-77: The shell step uses the GNU-only timeout binary ("timeout
600 claude -p --model claude-sonnet-4-6 <<< "$PROMPT"") which breaks on macOS;
replace it with a portable wrapper that picks timeout or gtimeout, e.g. set
TIMEOUT_BIN="$(command -v timeout || command -v gtimeout)" and invoke
"$TIMEOUT_BIN" 600 claude -p --model claude-sonnet-4-6 <<< "$PROMPT"; apply the
same change to the analogous invocation in catchup.md.In
@claws/openclaw/sageox/scripts/install-ox-curl.sh:
- Around line 40-42: Preflight currently checks for curl and tar but not jq,
causing late failures; add the same preflight check used for curl/tar to verify
jq is installed and exit with code 3 and an error message if missing; apply this
change both in the initial preflight block (the command -v curl/tar checks) and
the later install/check block around lines referenced (the block handling
state-file emission), so the script fails early with the documented tool-missing
flow when jq is absent.In
@claws/openclaw/sageox/scripts/update-ox.sh:
- Line 31: Update the remediation path strings in the echo messages inside
update-ox.sh to point to the correct doc (references/setup.md) instead of
references/INSTALL.md; locate the echo lines that print "ox install state not
configured — run install flow from references/INSTALL.md" (and the similar
messages at the other occurrences) and replace the INSTALL.md path with setup.md
so the error messages match the SKILL.md prerequisite flow.In
@claws/openclaw/sageox/scripts/update-state.sh:
- Around line 88-91: Replace the predictable TMP_FILE assignment that uses the
PID suffix with a secure mktemp call: create a temp file in the same directory
as STATE_FILE using mktemp, assign it to TMP_FILE, write NEW_STATE into
TMP_FILE, and then atomically mv it to STATE_FILE; ensure the trap cleans up the
mktemp-created TMP_FILE and that failure paths still remove TMP_FILE before
exit. Reference the TMP_FILE, STATE_FILE, NEW_STATE variables and the existing
trap/mv logic when making the change.In
@claws/openclaw/sageox/SKILL.md:
- Around line 50-52: Update the prerequisite flow in SKILL.md to include an
explicit "ox agent prime" step before any other SageOx capability checks or
actions (specifically insert this before the sentence starting "Before doing
anything else, verify the environment. Run every check in order." and the
repeated block around lines 75-82); state that you must runox agent prime
first and stop if it hasn't been executed, and adjust the explanatory text to
mention that this step enables SageOx queries and is required even if
install/auth checks pass.In
@claws/openclaw/sageox/test-skill.py:
- Around line 93-97: Replace the insecure SSH option in the command argument
list: where the SSH invocation is built (the list containing "ssh", "-i",
str(VPS_KEY), "-o", "StrictHostKeyChecking=no", "-o", "ConnectTimeout=10",
VPS_HOST), change the "-o", "StrictHostKeyChecking=no" element to use "-o",
"StrictHostKeyChecking=accept-new" (or remove it and point SSH at a managed
known_hosts file via "-o", "UserKnownHostsFile=/path/to/known_hosts") so tests
do not silently accept MITM; keep VPS_KEY and VPS_HOST usage intact.- Line 31: The constant OUTPUT_DIR is set to a shared /tmp path which can expose
sensitive test outputs; change the default to a private directory under the
current user's $HOME (e.g., Path(os.environ.get("HOME", "~")) /
".sageox-test-results") and ensure the code that creates the directory (the
place where OUTPUT_DIR is used/created, and the other occurrences around the
second occurrence at lines referenced 126-127) sets restrictive permissions
(mode 0o700) and uses mkdir(parents=True, exist_ok=True) so only the user can
read/write the output; update any related tests or docstrings to mention the new
default and keep the original /tmp behavior only if an explicit environment
variable or config overrides OUTPUT_DIR.- Around line 92-99: The ssh_cmd construction uses repr(prompt) which can lead
to unsafe remote shell interpretation; replace repr(prompt) with
shlex.quote(prompt) and also wrap session_id with shlex.quote(session_id) so
both values are safely shell-escaped, update the f-string in ssh_cmd to use
those quoted values, and add an import for shlex at the top of the file; locate
usage in the ssh_cmd list (and any other places using session_id or prompt for
remote shell commands) and apply the same quoting.
Nitpick comments:
In@claws/openclaw/sageox/deploy-to-vps.sh:
- Line 8: The SKILL_DIR variable in deploy-to-vps.sh uses a quoted tilde
("~/.openclaw/...") which works but is less explicit; change SKILL_DIR to use
$HOME (e.g., SKILL_DIR="$HOME/.openclaw/workspace/skills/sageox") for clarity
and consistency with VPS_KEY and other env vars, ensuring the path expands
correctly in local and remote contexts (refer to the SKILL_DIR assignment in
deploy-to-vps.sh).In
@claws/openclaw/sageox/references/setup.md:
- Around line 72-74: Add a concrete verification command for Claude credentials:
update the checklist step mentioningclaude/ANTHROPIC_API_KEYto include a
shell check that verifies theclaudeCLI is available and that either
~/.claude/config.jsonexists or theANTHROPIC_API_KEYenvironment variable
is set; if the check fails, instruct to runclaude loginor export
ANTHROPIC_API_KEYbefore launching OpenClaw.</details> <details> <summary>🪄 Autofix (Beta)</summary> Fix all unresolved CodeRabbit comments on this PR: - [ ] <!-- {"checkboxId": "4b0d0e0a-96d7-4f10-b296-3a18ea78f0b9"} --> Push a commit to this branch (recommended) - [ ] <!-- {"checkboxId": "ff5b1114-7d8c-49e6-8ac1-43f82af23a33"} --> Create a new PR with the fixes </details> --- <details> <summary>ℹ️ Review info</summary> <details> <summary>⚙️ Run configuration</summary> **Configuration used**: defaults **Review profile**: CHILL **Plan**: Pro Plus **Run ID**: `67d56bcf-a10d-46d2-bb58-732265a1b814` </details> <details> <summary>📥 Commits</summary> Reviewing files that changed from the base of the PR and between 760aeb03a4452e98b473a49dcc1e3a7e481e0ce2 and ea8267c462d820378ea97cc4e9e26509b257b8f7. </details> <details> <summary>📒 Files selected for processing (20)</summary> * `claws/openclaw/sageox/.clawhubignore` * `claws/openclaw/sageox/README.md` * `claws/openclaw/sageox/SKILL.md` * `claws/openclaw/sageox/assets/CATCHUP.md` * `claws/openclaw/sageox/assets/SUMMARIZE.md` * `claws/openclaw/sageox/deploy-to-vps.sh` * `claws/openclaw/sageox/references/catchup.md` * `claws/openclaw/sageox/references/coworkers.md` * `claws/openclaw/sageox/references/distill-pipeline.md` * `claws/openclaw/sageox/references/distill.md` * `claws/openclaw/sageox/references/glance.md` * `claws/openclaw/sageox/references/import-export.md` * `claws/openclaw/sageox/references/query.md` * `claws/openclaw/sageox/references/setup.md` * `claws/openclaw/sageox/references/summary.md` * `claws/openclaw/sageox/scripts/install-ox-curl.sh` * `claws/openclaw/sageox/scripts/select-new-entries.sh` * `claws/openclaw/sageox/scripts/update-ox.sh` * `claws/openclaw/sageox/scripts/update-state.sh` * `claws/openclaw/sageox/test-skill.py` </details> </details> <!-- This is an auto-generated comment by CodeRabbit for review status -->
| ## Prerequisites | ||
|
|
||
| - `ox` — installed via bundled curl script (not Homebrew) | ||
| - `git`, `gh`, `jq`, `claude` — declared in skill metadata | ||
|
|
There was a problem hiding this comment.
Prerequisites are missing required ox bootstrap/prime steps.
Please add ox login, ox init, ox doctor, ox status, and ox agent prime; without these, users can install the skill but still fail core SageOx query flows.
Based on learnings: Run ox agent prime before ANY other action to enable SageOx queries; initialize with ox login, ox init, ox doctor, then verify via ox status.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@claws/openclaw/sageox/README.md` around lines 18 - 22, Update the
Prerequisites section to include the missing ox bootstrap/prime steps:
explicitly list and order the commands users must run (start with `ox agent
prime` first, then `ox login`, `ox init`, `ox doctor`, and finally `ox status`)
and add a short note that `ox agent prime` must be run before any other action
to enable SageOx queries; ensure these commands are placed alongside the
existing `ox` installation note so users see installation + required
initialization steps together.
| Before doing anything else, verify the environment. Run every check in | ||
| order. If any fails, explain what's missing and stop. | ||
|
|
There was a problem hiding this comment.
Add an explicit ox agent prime prerequisite before capability execution.
The prerequisite flow omits ox agent prime, which is required in this repo before SageOx actions. Without it, later commands can fail despite passing install/auth checks.
Suggested patch
### 3. Authentication
1. `ox status` — confirm authenticated. If not: `ox login`.
2. `gh auth status` — confirm GitHub credentials.
3. `git config user.name` — confirm git identity.
4. `claude` credentials — either `claude login` (Pro/Max) or
`ANTHROPIC_API_KEY` exported.
+5. `ox agent prime` — run before any SageOx operation.Based on learnings: Run ox agent prime before ANY other action to enable SageOx queries.
Also applies to: 75-82
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@claws/openclaw/sageox/SKILL.md` around lines 50 - 52, Update the prerequisite
flow in SKILL.md to include an explicit "ox agent prime" step before any other
SageOx capability checks or actions (specifically insert this before the
sentence starting "Before doing anything else, verify the environment. Run every
check in order." and the repeated block around lines 75-82); state that you must
run `ox agent prime` first and stop if it hasn't been executed, and adjust the
explanatory text to mention that this step enables SageOx queries and is
required even if install/auth checks pass.
- Fix INSTALL.md -> setup.md path mismatch in update-ox.sh - Add jq preflight check in install-ox-curl.sh - Use mktemp instead of predictable .$$ in update-state.sh - Make timeout portable (timeout/gtimeout) in summary.md and catchup.md - Add MD040 language tags to fenced code blocks in summary.md - Use shlex.quote instead of repr() for shell escaping in test-skill.py - Use StrictHostKeyChecking=accept-new instead of =no - Move test output to ~/.sageox-test-results with 0o700 permissions Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-Authored-By: SageOx <ox@sageox.ai>
There was a problem hiding this comment.
Actionable comments posted: 5
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@claws/openclaw/sageox/references/catchup.md`:
- Around line 49-50: The documentation references the template as
`assets/CATCHUP.md` which is ambiguous because step 2 runs in the selected repo
directory; update the reference to use an unambiguous, absolute or
skill-root-relative template path (e.g.,
`openclaw/skills/<skill-name>/assets/CATCHUP.md` or a full repo-root path)
wherever the prompt is built (the "Build a prompt by substituting gathered data
into `assets/CATCHUP.md`" step) so that the templating logic loads the correct
`CATCHUP.md` from the OpenClaw skill root rather than the selected repo
directory.
- Around line 58-60: The TIMEOUT_BIN assignment and usage
(TIMEOUT_BIN="$(command -v timeout || command -v gtimeout)" and "$TIMEOUT_BIN"
600 claude ...) must be guarded: detect when TIMEOUT_BIN is empty before
invoking it and emit a clear error/exit (e.g., print "timeout or gtimeout not
found" and exit 1) so the claude call isn't attempted with an empty command;
apply the same guard to the identical pattern in summary.md (the TIMEOUT_BIN
variable and its invocation) to avoid non-actionable failures.
In `@claws/openclaw/sageox/scripts/install-ox-curl.sh`:
- Around line 207-218: The TMP_STATE_FILE created by mktemp can be left behind
if jq fails because the existing EXIT trap only removes WORK_DIR; initialize
TMP_STATE_FILE="" before the trap and extend the EXIT trap handler (the same
trap declared around WORK_DIR) to also check and remove TMP_STATE_FILE if
non-empty, ensuring cleanup even on set -euo pipefail failures; locate the
mktemp/TMP_STATE_FILE usage, the jq -> mv block, and the existing trap (the
handler that currently cleans WORK_DIR) to update it accordingly and avoid
adding a separate trap.
In `@claws/openclaw/sageox/test-skill.py`:
- Line 127: The current call OUTPUT_DIR.mkdir(parents=True, exist_ok=True,
mode=0o700) only sets permissions for newly created directories and leaves
pre-existing ones unchanged; after the mkdir call, enforce the desired
permissions by applying a chmod to OUTPUT_DIR (e.g., call os.chmod(OUTPUT_DIR,
0o700) or OUTPUT_DIR.chmod(0o700)) so existing directories get the correct 0o700
mode, and wrap it in a try/except if you need to surface or log permission
errors.
- Around line 131-134: The SSH preflight subprocess.run invocation omits the
StrictHostKeyChecking=accept-new option causing failures on fresh hosts; update
the argument list in the preflight check (the subprocess.run call creating
"check") to include "-o", "StrictHostKeyChecking=accept-new" (matching the
option used by send_prompt) so the preflight behaves the same as send_prompt and
won't hang or prompt for interactive confirmation.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro Plus
Run ID: 40190353-601a-4f9c-af81-3d0346f496b3
📒 Files selected for processing (6)
claws/openclaw/sageox/references/catchup.mdclaws/openclaw/sageox/references/summary.mdclaws/openclaw/sageox/scripts/install-ox-curl.shclaws/openclaw/sageox/scripts/update-ox.shclaws/openclaw/sageox/scripts/update-state.shclaws/openclaw/sageox/test-skill.py
✅ Files skipped from review due to trivial changes (1)
- claws/openclaw/sageox/references/summary.md
🚧 Files skipped from review as they are similar to previous changes (1)
- claws/openclaw/sageox/scripts/update-state.sh
| Build a prompt by substituting gathered data into `assets/CATCHUP.md`: | ||
| - `{{DISTILLED}}` — distilled daily entries (content from step 2) |
There was a problem hiding this comment.
Use an unambiguous template path for CATCHUP.md.
Step 2 explicitly runs from the selected repo directory, but this step references assets/CATCHUP.md as if it were local to that repo. That can break prompt templating when the skill assets live under the OpenClaw skill root.
Suggested doc fix
-Build a prompt by substituting gathered data into `assets/CATCHUP.md`:
+Build a prompt by substituting gathered data into
+`claws/openclaw/sageox/assets/CATCHUP.md` (or an equivalent resolved skill-root path):📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| Build a prompt by substituting gathered data into `assets/CATCHUP.md`: | |
| - `{{DISTILLED}}` — distilled daily entries (content from step 2) | |
| Build a prompt by substituting gathered data into | |
| `claws/openclaw/sageox/assets/CATCHUP.md` (or an equivalent resolved skill-root path): | |
| - `{{DISTILLED}}` — distilled daily entries (content from step 2) |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@claws/openclaw/sageox/references/catchup.md` around lines 49 - 50, The
documentation references the template as `assets/CATCHUP.md` which is ambiguous
because step 2 runs in the selected repo directory; update the reference to use
an unambiguous, absolute or skill-root-relative template path (e.g.,
`openclaw/skills/<skill-name>/assets/CATCHUP.md` or a full repo-root path)
wherever the prompt is built (the "Build a prompt by substituting gathered data
into `assets/CATCHUP.md`" step) so that the templating logic loads the correct
`CATCHUP.md` from the OpenClaw skill root rather than the selected repo
directory.
| TIMEOUT_BIN="$(command -v timeout || command -v gtimeout)" | ||
| "$TIMEOUT_BIN" 600 claude -p --model claude-sonnet-4-6 <<< "$PROMPT" | ||
| ``` |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Verify whether docs/scripts in the repo already guarantee timeout availability.
# Expected:
# - Either a documented prerequisite mentioning timeout/gtimeout
# - Or script-level preflight checks before timeout execution
rg -n -C2 'command -v timeout|gtimeout|TIMEOUT_BIN|coreutils|prereq|preflight' claws openclaw || trueRepository: sageox/ox
Length of output: 3287
🏁 Script executed:
# First, let's check the full context of catchup.md around the timeout usage
cat -n claws/openclaw/sageox/references/catchup.md | head -80 | tail -30Repository: sageox/ox
Length of output: 1291
🏁 Script executed:
# Check if there's documentation about prerequisites/requirements
git ls-files | grep -i -E '(readme|prerequisite|setup|install|requirement)' | head -20Repository: sageox/ox
Length of output: 745
🏁 Script executed:
# Search for any preflight or prerequisite validation scripts
rg -n 'prerequisite|preflight|check.*timeout|check.*timeout' --type md --type sh --type bashRepository: sageox/ox
Length of output: 82
🏁 Script executed:
# Check setup.md and INSTALL.md for timeout/gtimeout prerequisites
cat claws/openclaw/sageox/references/setup.md | head -100Repository: sageox/ox
Length of output: 3186
🏁 Script executed:
# Check the main README for timeout mention
rg -n -A2 -B2 'timeout|gtimeout|coreutils|GNU|prerequisite' claws/openclaw/sageox/README.mdRepository: sageox/ox
Length of output: 35
🏁 Script executed:
# Search for any validation or check for timeout in the skill code
rg -n 'timeout|gtimeout' claws/openclaw/sageox/scripts/ claws/openclaw/sageox/references/Repository: sageox/ox
Length of output: 504
🏁 Script executed:
# Check if there's any error handling in claude command invocation
rg -n -A3 'TIMEOUT_BIN.*600|claude.*600' claws/openclaw/sageox/Repository: sageox/ox
Length of output: 719
Add guard for missing timeout/gtimeout binary before synthesis execution.
The code assumes timeout (or gtimeout on macOS) is available but doesn't validate it exists. If neither binary is present, TIMEOUT_BIN becomes empty and the claude command fails with a non-actionable error. Add an explicit check before execution.
This same pattern appears in summary.md (lines 76–77) and should also be fixed there.
Suggested fix
TIMEOUT_BIN="$(command -v timeout || command -v gtimeout)"
+if [ -z "$TIMEOUT_BIN" ]; then
+ echo "Error: timeout utility not found. Install coreutils (timeout) or switch to macOS (gtimeout)."
+ return 1
+fi
"$TIMEOUT_BIN" 600 claude -p --model claude-sonnet-4-6 <<< "$PROMPT"📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| TIMEOUT_BIN="$(command -v timeout || command -v gtimeout)" | |
| "$TIMEOUT_BIN" 600 claude -p --model claude-sonnet-4-6 <<< "$PROMPT" | |
| ``` | |
| TIMEOUT_BIN="$(command -v timeout || command -v gtimeout)" | |
| if [ -z "$TIMEOUT_BIN" ]; then | |
| echo "Error: timeout utility not found. Install coreutils (timeout) or switch to macOS (gtimeout)." | |
| return 1 | |
| fi | |
| "$TIMEOUT_BIN" 600 claude -p --model claude-sonnet-4-6 <<< "$PROMPT" |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@claws/openclaw/sageox/references/catchup.md` around lines 58 - 60, The
TIMEOUT_BIN assignment and usage (TIMEOUT_BIN="$(command -v timeout || command
-v gtimeout)" and "$TIMEOUT_BIN" 600 claude ...) must be guarded: detect when
TIMEOUT_BIN is empty before invoking it and emit a clear error/exit (e.g., print
"timeout or gtimeout not found" and exit 1) so the claude call isn't attempted
with an empty command; apply the same guard to the identical pattern in
summary.md (the TIMEOUT_BIN variable and its invocation) to avoid non-actionable
failures.
- Add TMP_STATE_FILE cleanup to EXIT trap in install-ox-curl.sh - Enforce 0o700 on existing output dir with os.chmod in test-skill.py - Add StrictHostKeyChecking=accept-new to SSH preflight check Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-Authored-By: SageOx <ox@sageox.ai>
1 participant
Summary
sageoxOpenClaw skill with 8 capabilities, replacing the separatesageox-distillandsageox-summaryskillssageox-distill-repos.json) as central context anchor — all ox commandscdto a manifest repo before runningCapabilities
ox queryox coworkerox distillclaude -pox glanceTest plan
openclaw agent --session-id— 10/10 passingCo-Authored-By: SageOx
🤖 Generated with Claude Code
Summary by CodeRabbit