feat: add usage statistics persistence for cliproxyapi - #674
Conversation
|
You do not have enough credits to review this pull request. Please purchase more credits to continue. |
Summary of ChangesHello @shunkakinoki, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request enhances the cliproxyapi service by introducing robust usage statistics persistence. It ensures that usage data is automatically saved when the service shuts down and restored upon startup, preventing data loss. Additionally, it refactors the Docker image pulling process into the service's startup script and implements graceful shutdown procedures to maintain data integrity. Highlights
Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here. You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension. Footnotes
|
|
Note Other AI code review bot(s) detectedCodeRabbit has detected other AI code review bot(s) in this pull request and will avoid duplicating their findings in the review comments. This may lead to a less comprehensive review. 📝 WalkthroughSummary by CodeRabbit
✏️ Tip: You can customize this high-level summary in your review settings. WalkthroughAdds usage-statistics import/export and management polling to the cliproxyapi startup script; reworks background startup and coordinated shutdown for containers/binaries. Removes docker pull logic from the Makefile's systemctl-cliproxyapi target. Documentation updated to reflect service-oriented startup and usage handling. Changes
Sequence Diagram(s)sequenceDiagram
participant Script as start.sh (script)
participant Host as Docker / Homebrew (host runtime)
participant Management as Management Endpoint
Script->>Host: Launch container or binary in background
Note over Script: Record child PID
Script->>Management: Poll readiness (wait_for_management)
loop Poll until ready or timeout
Management-->>Script: Health/ready response
end
Script->>Management: POST usage data (usage_import)
Management-->>Script: Acknowledge import
Script->>Host: Wait for child process to exit
Host-->>Script: Child process exits (status)
Script->>Management: GET/export usage data (usage_export)
Management-->>Script: Return usage payload (saved to file)
Script->>Script: Cleanup and exit with child status
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 2 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing touches
🧪 Generate unit tests (beta)
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 |
Mesa DescriptionTL;DRPersisted What changed?
Description generated by Mesa. Update settings |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces a valuable feature for persisting usage statistics across service restarts. The approach of using trap handlers for graceful shutdown and curl for interacting with the management API is sound. The refactoring of the Docker pull logic into start.sh is also a good improvement for centralizing service startup logic.
My review has identified a few issues in the start.sh script. There is a critical issue in how the management API key is handled, which could lead to authentication failures. I've also found a high-severity issue with the EXIT trap that would cause it to fail on every normal service exit. Additionally, there is an opportunity to improve maintainability by refactoring duplicated code blocks. Addressing these points will make the implementation more robust and easier to maintain.
| MANAGEMENT_PASSWORD="${CLIPROXY_MANAGEMENT_PASSWORD:-}" | ||
| MANAGEMENT_KEY="${CLIPROXY_MANAGEMENT_PASSWORD:-${CLIPROXY_MANAGEMENT_KEY:-}}" |
There was a problem hiding this comment.
There's a potential bug in how MANAGEMENT_PASSWORD and MANAGEMENT_KEY are derived. If CLIPROXY_MANAGEMENT_KEY is set but CLIPROXY_MANAGEMENT_PASSWORD is not, MANAGEMENT_PASSWORD will be empty. This empty password is then used to generate config.yaml (via sed on line 79) and passed as an environment variable to the Docker container. However, MANAGEMENT_KEY will hold the value of CLIPROXY_MANAGEMENT_KEY, and this script will use it to make API calls, which will fail due to the password mismatch.
To fix this, the service must be configured with the correct key, and the script must use that same key. The suggested change unifies the key derivation. Note: For this to work completely, you must also update the sed command on line 79 to use ${MANAGEMENT_PASSWORD} instead of ${CLIPROXY_MANAGEMENT_PASSWORD:-}.
| MANAGEMENT_PASSWORD="${CLIPROXY_MANAGEMENT_PASSWORD:-}" | |
| MANAGEMENT_KEY="${CLIPROXY_MANAGEMENT_PASSWORD:-${CLIPROXY_MANAGEMENT_KEY:-}}" | |
| MANAGEMENT_KEY="${CLIPROXY_MANAGEMENT_KEY:-${CLIPROXY_MANAGEMENT_PASSWORD:-}}" | |
| MANAGEMENT_PASSWORD="${MANAGEMENT_KEY}" |
| } | ||
|
|
||
| child_pid="" | ||
| trap 'usage_export' EXIT |
There was a problem hiding this comment.
The EXIT trap as currently implemented will cause issues. The EXIT trap is executed when the script exits. In the normal flow of this script, it exits after the child process (the cliproxyapi service) has terminated. At that point, the service is no longer running, so the usage_export call will fail to connect to the management API.
This trap is redundant because the TERM/INT trap already handles exporting usage statistics before terminating the service in graceful shutdown scenarios. For abnormal exits (like the service crashing), this EXIT trap wouldn't be able to save the stats before the service is gone.
I recommend removing the EXIT trap to avoid failed API calls on every normal service exit.
| child_pid=$! | ||
| wait_for_management || true | ||
| usage_import | ||
| wait "$child_pid" | ||
| exit $? |
There was a problem hiding this comment.
This block of logic for setting the child_pid, waiting for the management API, importing usage, and waiting for the process is duplicated for the macOS execution paths (lines 212-216 and 219-223). This code repetition makes the script harder to maintain.
To improve this, you could extract the common logic into a helper function. For example, you could define this function before the platform-specific if blocks:
run_and_wait() {
child_pid=$!
wait_for_management || true
usage_import
wait "$child_pid"
exit $?
}Then, you can replace this block and the other duplicated blocks with a single call: run_and_wait.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Fix all issues with AI agents
In `@home-manager/services/cliproxyapi/scripts/start.sh`:
- Around line 179-181: The TERM/INT trap redundantly calls usage_export and then
the EXIT trap runs it again; update the TERM/INT trap to disable the EXIT trap
first (e.g., run trap - EXIT) before invoking usage_export and terminating the
child process so usage_export only runs once; reference the existing trap that
mentions usage_export, child_pid, and the TERM/INT trap and ensure you preserve
the kill/wait logic and redirecting errors while removing the duplicate EXIT
invocation.
- Around line 161-177: The readiness check in wait_for_management uses curl but
doesn't treat HTTP 4xx/5xx as failures; update the curl invocation inside
wait_for_management to include --fail so non-2xx responses cause curl to return
a non-zero exit code (keep existing -sS, -H "Authorization: Bearer
${MANAGEMENT_KEY}" and -o /dev/null options intact) so the loop will retry on
unauthorized or server errors.
🧹 Nitpick comments (1)
home-manager/services/cliproxyapi/scripts/start.sh (1)
210-223: Consider extracting duplicated startup logic.The startup sequence (background exec, PID tracking, wait, import, wait, exit) is identical for both Homebrew paths. This could be refactored to reduce duplication, though it's a minor concern given the script's scope.
Optional refactor
+run_binary() { + local binary="$1" + shift + "$binary" -config "$CONFIG" "$@" & + child_pid=$! + wait_for_management || true + usage_import + wait "$child_pid" + exit $? +} + # macOS: Homebrew binary if [ -x /opt/homebrew/bin/cliproxyapi ]; then - /opt/homebrew/bin/cliproxyapi -config "$CONFIG" "$@" & - child_pid=$! - wait_for_management || true - usage_import - wait "$child_pid" - exit $? + run_binary /opt/homebrew/bin/cliproxyapi "$@" elif [ -x /usr/local/bin/cliproxyapi ]; then - /usr/local/bin/cliproxyapi -config "$CONFIG" "$@" & - child_pid=$! - wait_for_management || true - usage_import - wait "$child_pid" - exit $? + run_binary /usr/local/bin/cliproxyapi "$@" else
| wait_for_management() { | ||
| if [ -z "$MANAGEMENT_KEY" ]; then | ||
| return 0 | ||
| fi | ||
| local attempts=60 | ||
| local i | ||
| for i in $(seq 1 "$attempts"); do | ||
| if curl -sS \ | ||
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | ||
| "${MANAGEMENT_URL}/usage/export" \ | ||
| -o /dev/null; then | ||
| return 0 | ||
| fi | ||
| sleep 3 | ||
| done | ||
| return 1 | ||
| } |
There was a problem hiding this comment.
Add --fail to validate HTTP status in readiness check.
The curl command returns exit code 0 for any HTTP response, including 4xx/5xx errors. This means a 401 Unauthorized or 500 Server Error would be incorrectly treated as "ready."
Proposed fix
if curl -sS \
+ --fail \
-H "Authorization: Bearer ${MANAGEMENT_KEY}" \
"${MANAGEMENT_URL}/usage/export" \
-o /dev/null; then📝 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.
| wait_for_management() { | |
| if [ -z "$MANAGEMENT_KEY" ]; then | |
| return 0 | |
| fi | |
| local attempts=60 | |
| local i | |
| for i in $(seq 1 "$attempts"); do | |
| if curl -sS \ | |
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | |
| "${MANAGEMENT_URL}/usage/export" \ | |
| -o /dev/null; then | |
| return 0 | |
| fi | |
| sleep 3 | |
| done | |
| return 1 | |
| } | |
| wait_for_management() { | |
| if [ -z "$MANAGEMENT_KEY" ]; then | |
| return 0 | |
| fi | |
| local attempts=60 | |
| local i | |
| for i in $(seq 1 "$attempts"); do | |
| if curl -sS \ | |
| --fail \ | |
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | |
| "${MANAGEMENT_URL}/usage/export" \ | |
| -o /dev/null; then | |
| return 0 | |
| fi | |
| sleep 3 | |
| done | |
| return 1 | |
| } |
🤖 Prompt for AI Agents
In `@home-manager/services/cliproxyapi/scripts/start.sh` around lines 161 - 177,
The readiness check in wait_for_management uses curl but doesn't treat HTTP
4xx/5xx as failures; update the curl invocation inside wait_for_management to
include --fail so non-2xx responses cause curl to return a non-zero exit code
(keep existing -sS, -H "Authorization: Bearer ${MANAGEMENT_KEY}" and -o
/dev/null options intact) so the loop will retry on unauthorized or server
errors.
| child_pid="" | ||
| trap 'usage_export' EXIT | ||
| trap 'usage_export; if [ -n "$child_pid" ]; then kill -TERM "$child_pid" 2>/dev/null || true; wait "$child_pid" 2>/dev/null || true; fi' TERM INT |
There was a problem hiding this comment.
Double usage_export on signal termination.
When receiving TERM/INT, the signal trap runs usage_export, then when the script exits, the EXIT trap also runs usage_export again. This is redundant and the second call may fail if the service is already down.
Proposed fix: Clear EXIT trap after signal handler runs
child_pid=""
trap 'usage_export' EXIT
-trap 'usage_export; if [ -n "$child_pid" ]; then kill -TERM "$child_pid" 2>/dev/null || true; wait "$child_pid" 2>/dev/null || true; fi' TERM INT
+trap 'trap - EXIT; usage_export; if [ -n "$child_pid" ]; then kill -TERM "$child_pid" 2>/dev/null || true; wait "$child_pid" 2>/dev/null || true; fi' TERM INT📝 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.
| child_pid="" | |
| trap 'usage_export' EXIT | |
| trap 'usage_export; if [ -n "$child_pid" ]; then kill -TERM "$child_pid" 2>/dev/null || true; wait "$child_pid" 2>/dev/null || true; fi' TERM INT | |
| child_pid="" | |
| trap 'usage_export' EXIT | |
| trap 'trap - EXIT; usage_export; if [ -n "$child_pid" ]; then kill -TERM "$child_pid" 2>/dev/null || true; wait "$child_pid" 2>/dev/null || true; fi' TERM INT |
🤖 Prompt for AI Agents
In `@home-manager/services/cliproxyapi/scripts/start.sh` around lines 179 - 181,
The TERM/INT trap redundantly calls usage_export and then the EXIT trap runs it
again; update the TERM/INT trap to disable the EXIT trap first (e.g., run trap -
EXIT) before invoking usage_export and terminating the child process so
usage_export only runs once; reference the existing trap that mentions
usage_export, child_pid, and the TERM/INT trap and ensure you preserve the
kill/wait logic and redirecting errors while removing the duplicate EXIT
invocation.
There was a problem hiding this comment.
Pull request overview
Adds automatic persistence of cliproxyapi usage statistics by exporting on shutdown and importing on startup, and relocates Docker image pull logic from the Makefile into the service start script.
Changes:
- Add usage stats export/import + management readiness wait + signal/exit traps in
start.sh. - Update service docs to reflect new start behavior.
- Remove Docker pull from
systemctl-cliproxyapiMakefile target.
Reviewed changes
Copilot reviewed 3 out of 3 changed files in this pull request and generated 9 comments.
| File | Description |
|---|---|
| home-manager/services/cliproxyapi/scripts/start.sh | Implements usage export/import persistence and moves Docker pull into startup flow. |
| home-manager/services/cliproxyapi/README.md | Updates script purpose wording and notes automatic usage backup/restore. |
| Makefile | Removes Docker pull step from the cliproxyapi systemctl target. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| curl -sS \ | ||
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | ||
| "${MANAGEMENT_URL}/usage/export" \ | ||
| -o "$USAGE_EXPORT_FILE" || true |
There was a problem hiding this comment.
usage-export.json is written without explicitly restricting file permissions. Depending on the user’s umask, this file could be group/world-readable and may contain sensitive usage metadata. Consider setting a restrictive umask for the export write or chmod 600 the file after a successful export.
| -o "$USAGE_EXPORT_FILE" || true | |
| -o "$USAGE_EXPORT_FILE" && chmod 600 "$USAGE_EXPORT_FILE" || true |
| usage_import() { | ||
| if [ -z "$MANAGEMENT_KEY" ] || [ ! -f "$USAGE_EXPORT_FILE" ]; then | ||
| return 0 | ||
| fi | ||
| curl -sS \ | ||
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | ||
| -H "Content-Type: application/json" \ | ||
| -X POST \ | ||
| --data-binary @"$USAGE_EXPORT_FILE" \ | ||
| "${MANAGEMENT_URL}/usage/import" >/dev/null || true | ||
| } | ||
|
|
||
| usage_export() { | ||
| if [ -z "$MANAGEMENT_KEY" ]; then | ||
| return 0 | ||
| fi | ||
| mkdir -p "$(dirname "$USAGE_EXPORT_FILE")" | ||
| curl -sS \ | ||
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | ||
| "${MANAGEMENT_URL}/usage/export" \ | ||
| -o "$USAGE_EXPORT_FILE" || true | ||
| } | ||
|
|
||
| wait_for_management() { | ||
| if [ -z "$MANAGEMENT_KEY" ]; then | ||
| return 0 | ||
| fi | ||
| local attempts=60 | ||
| local i | ||
| for i in $(seq 1 "$attempts"); do | ||
| if curl -sS \ | ||
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | ||
| "${MANAGEMENT_URL}/usage/export" \ | ||
| -o /dev/null; then | ||
| return 0 | ||
| fi | ||
| sleep 3 | ||
| done | ||
| return 1 | ||
| } | ||
|
|
||
| child_pid="" | ||
| trap 'usage_export' EXIT | ||
| trap 'usage_export; if [ -n "$child_pid" ]; then kill -TERM "$child_pid" 2>/dev/null || true; wait "$child_pid" 2>/dev/null || true; fi' TERM INT |
There was a problem hiding this comment.
This PR adds non-trivial new startup/shutdown behavior (usage import/export, readiness polling, traps), but the existing ShellSpec coverage for start.sh doesn’t exercise these paths. Adding tests that mock curl and verify (1) no calls happen without a management key, (2) import happens after readiness, and (3) export runs on exit would help prevent regressions.
| @@ -704,12 +704,6 @@ systemctl: systemctl-cliproxyapi systemctl-clawdbot systemctl-code-syncer system | |||
|
|
|||
| .PHONY: systemctl-cliproxyapi | |||
| systemctl-cliproxyapi: ## Pull latest image and restart cliproxyapi systemd user service. | |||
There was a problem hiding this comment.
The target description still says it pulls the latest image, but the pull logic was removed from this recipe. Update the help text to match current behavior (or restore the pull step if the description is still intended).
| systemctl-cliproxyapi: ## Pull latest image and restart cliproxyapi systemd user service. | |
| systemctl-cliproxyapi: ## Restart cliproxyapi systemd user service. |
| # OAuth login | ||
| cliproxyapi --claude-login | ||
|
|
||
| # Usage statistics backup/restore happens automatically on start/exit |
There was a problem hiding this comment.
The README now states usage statistics backup/restore happens automatically, but in start.sh it is a no-op unless a management key is configured (and the Management API/usage stats are enabled). Consider clarifying the prerequisites here (e.g., requires CLIPROXY_MANAGEMENT_PASSWORD/key and Management API enabled).
| # Usage statistics backup/restore happens automatically on start/exit | |
| # Usage statistics backup/restore happens automatically on start/exit | |
| # (requires Management API with usage stats enabled and CLIPROXY_MANAGEMENT_PASSWORD/management key configured) |
| if curl -sS \ | ||
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | ||
| "${MANAGEMENT_URL}/usage/export" \ | ||
| -o /dev/null; then | ||
| return 0 |
There was a problem hiding this comment.
wait_for_management treats any HTTP response as success because curl returns exit code 0 for 4xx/5xx unless --fail/--fail-with-body is used. This can mark the management API as “ready” even when the key is invalid (e.g., 401), and then usage_import/usage_export will operate against an unauthorized endpoint. Consider checking the HTTP status explicitly (or using --fail-with-body) so only a 2xx response is treated as ready.
| curl -sS \ | ||
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | ||
| "${MANAGEMENT_URL}/usage/export" \ | ||
| -o "$USAGE_EXPORT_FILE" || true |
There was a problem hiding this comment.
usage_export writes directly to usage-export.json even when the request fails (and without curl --fail), which risks overwriting a good backup with an error page/JSON error response. A safer pattern is to require a successful HTTP status and write to a temp file, then atomically move it into place on success.
| curl -sS \ | |
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | |
| "${MANAGEMENT_URL}/usage/export" \ | |
| -o "$USAGE_EXPORT_FILE" || true | |
| local tmp_file="${USAGE_EXPORT_FILE}.tmp.$$" | |
| if curl -sS --fail \ | |
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | |
| "${MANAGEMENT_URL}/usage/export" \ | |
| -o "$tmp_file"; then | |
| mv "$tmp_file" "$USAGE_EXPORT_FILE" | |
| else | |
| rm -f "$tmp_file" || true | |
| fi |
| curl -sS \ | ||
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | ||
| -H "Content-Type: application/json" \ | ||
| -X POST \ | ||
| --data-binary @"$USAGE_EXPORT_FILE" \ | ||
| "${MANAGEMENT_URL}/usage/import" >/dev/null || true | ||
| } | ||
|
|
||
| usage_export() { | ||
| if [ -z "$MANAGEMENT_KEY" ]; then | ||
| return 0 | ||
| fi | ||
| mkdir -p "$(dirname "$USAGE_EXPORT_FILE")" | ||
| curl -sS \ | ||
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | ||
| "${MANAGEMENT_URL}/usage/export" \ | ||
| -o "$USAGE_EXPORT_FILE" || true |
There was a problem hiding this comment.
All curl calls here have no connect/overall timeouts. Since usage_export runs in EXIT/TERM/INT traps, a stalled TCP connect (e.g., misconfigured CLIPROXY_MANAGEMENT_URL) can block shutdown indefinitely. Consider adding --connect-timeout and --max-time (or similar) to each curl invocation.
| curl -sS \ | ||
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | ||
| -H "Content-Type: application/json" \ | ||
| -X POST \ | ||
| --data-binary @"$USAGE_EXPORT_FILE" \ | ||
| "${MANAGEMENT_URL}/usage/import" >/dev/null || true | ||
| } | ||
|
|
||
| usage_export() { | ||
| if [ -z "$MANAGEMENT_KEY" ]; then | ||
| return 0 | ||
| fi | ||
| mkdir -p "$(dirname "$USAGE_EXPORT_FILE")" | ||
| curl -sS \ | ||
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | ||
| "${MANAGEMENT_URL}/usage/export" \ | ||
| -o "$USAGE_EXPORT_FILE" || true | ||
| } | ||
|
|
||
| wait_for_management() { | ||
| if [ -z "$MANAGEMENT_KEY" ]; then | ||
| return 0 | ||
| fi | ||
| local attempts=60 | ||
| local i | ||
| for i in $(seq 1 "$attempts"); do | ||
| if curl -sS \ | ||
| -H "Authorization: Bearer ${MANAGEMENT_KEY}" \ | ||
| "${MANAGEMENT_URL}/usage/export" \ |
There was a problem hiding this comment.
The curl invocations include the MANAGEMENT_KEY directly in an Authorization header on the process command line, which can leak this secret via ps//proc to other local users or monitoring tooling. An attacker with local access could capture the bearer token while these commands run and reuse it to call the management API with full privileges. To mitigate this, avoid placing secrets in command-line arguments (e.g., pass the token via environment variables, a config file with restricted permissions, or curl config, and ensure it is not exposed in process listings).
| if docker info >/dev/null 2>&1; then | ||
| echo "🔄 Pulling latest cliproxyapi image..." | ||
| docker pull eceasy/cli-proxy-api:latest || true | ||
| else | ||
| echo "⏭️ Skipping docker pull (docker not accessible)" | ||
| fi | ||
|
|
||
| docker rm -f cliproxyapi 2>/dev/null || true | ||
| mkdir -p "$CONFIG_DIR/logs" | ||
| exec docker run --rm \ | ||
| docker run --rm \ | ||
| --name cliproxyapi \ | ||
| --network host \ | ||
| --ulimit nofile=65536:65536 \ | ||
| -v "$CONFIG:/CLIProxyAPI/config.yaml:ro" \ | ||
| -v "$CONFIG_DIR:/root/.cli-proxy-api" \ | ||
| -v "$CONFIG_DIR/logs:/CLIProxyAPI/logs" \ | ||
| -e MANAGEMENT_PASSWORD="${MANAGEMENT_PASSWORD:-}" \ | ||
| eceasy/cli-proxy-api:latest | ||
| eceasy/cli-proxy-api:latest & |
There was a problem hiding this comment.
This script pulls and runs the Docker image eceasy/cli-proxy-api:latest using a mutable latest tag from Docker Hub, which creates a supply chain risk if the upstream image is compromised or the tag is retagged maliciously. Because the container is started with access to configuration, auth directories, and management credentials, a poisoned image would have full control over the service and associated secrets. To reduce this risk, pin the image to an immutable reference (such as a specific version tag or image digest) and update it explicitly when upgrading.
![cubic-dev-ai[bot]](https://avatars.githubusercontent.com/in/1082092?s=60&v=4){kind=small}
Changes
Technical Details
Testing
Generated with opencode by glm-4.7
Summary by cubic
Persist cliproxyapi usage statistics across restarts by auto-exporting on shutdown and restoring on startup. Moved Docker image pull into start.sh and added graceful shutdown traps.
New Features
Refactors
Written for commit 00d610e. Summary will update on new commits.