Inspect CARLA driving simulation logs without replaying them. Get recording details, find collisions by participant category, and identify vehicles that barely moved, with times and object IDs to locate moments worth reviewing.
Inspects a recorded CARLA .log without replaying it — show_recorder_file_info for the header/actors/frames, collision queries between actor categories, and blocked-actor queries for vehicles that got stuck. Use when the user asks to "what's in this recording", "show recorder info", "which vehicles collided", "find collisions in the log", or "which actors were blocked/stuck".
Review the instructions the agent will receive before you install.
Paths.
scripts/…andreferences/…below are relative to the directory holding this SKILL.md. Your working directory is the user's project, not that directory, so prefix them with its absolute path or the command is not found.
Read facts out of a .log without replaying it. The server parses the file and
returns a text report. Three questions:
info (header: map, duration, frames, actor list).collisions between two actor categories.blocked actors that barely moved.Progress:
- [ ] Step 1: Check prerequisites (bash scripts/check_env.sh), clear FAILs
- [ ] Step 2: Run the query that answers the question (info / collisions / blocked)
- [ ] Step 3: Read the report; use ids/times to drive a targeted replay if needed
Commands need CARLA_HOST/CARLA_PORT from scripts/env.sh.
bash scripts/check_env.sh
source scripts/env.sh
# header + actors (add --all for every frame; large)
python3 scripts/query.py info --file /tmp/run.log
# collisions: category codes h=hero v=vehicle w=walker t=traffic-light o=other a=any
python3 scripts/query.py collisions --file /tmp/run.log --type1 v --type2 a # any vehicle vs anything
python3 scripts/query.py collisions --file /tmp/run.log --type1 v --type2 w # vehicle vs walker
# actors that moved < 1 m (100 cm) over >= 30 s (stuck)
python3 scripts/query.py blocked --file /tmp/run.log --min-time 30 --min-distance 100
Units matter: --min-time is seconds, --min-distance is centimetres
(100 = 1 m).
The reports give frames, times and actor ids — feed them to
replay-recording (--start, --follow) to
jump to and watch the interesting moment.
Example 1: overview
User says: "what's in /tmp/run.log?"
info --file /tmp/run.log → map, duration, frame count, actors.
Example 2: find a crash
User says: "did any car hit a pedestrian?"
collisions --file /tmp/run.log --type1 v --type2 w. Each line gives the frame,
time and the two actor ids; then replay play --start <t> --follow <id>.
Example 3: find gridlock
User says: "which cars got stuck?"
blocked --file /tmp/run.log --min-time 20 --min-distance 50 (moved < 0.5 m in
20 s).
Problem: parse error / empty report
Cause: the .log is not on the server at that path, or has no frames.
Solution: use the server-side path; confirm the recording with info first.
Problem: collisions returns nothing but I saw a crash
Cause: wrong categories, or collisions need additional_data context.
Solution: widen with --type2 a (any); ensure the recording captured the actors.
Problem: blocked lists everything / nothing
Cause: --min-distance unit confusion (centimetres, not metres).
Solution: 100 = 1 m; raise --min-time to ignore brief stops at lights.
Text reports printed to stdout (no world change, no file written). Use them to
target a replay-recording.
Detail (report fields, category codes, blocked-query semantics) in references/queries.md.
No reviews yet. Tell others what you thought.
No login needed. Share how it worked for you, ask a question, or note where it did not run.
No comments yet. Be the first to write one.
Add a message type to CARLA's native ROS 2 interface. Create its C++ data structure, serialization code for transmission, type identity hash, registration, and tests, with Docker-based hash calculation and build checks.
Extend CARLA’s native ROS 2 interface to publish sensor data or receive vehicle commands. Covers topic implementation, registration, data connections, and verification, plus adding middleware, the communication layer used to exchange messages.
Summarize ScenarioRunner results to see which scenarios passed and which checks failed. Compare runs before and after changes, or analyze recorded driving to produce measurements and plots of vehicle distance and deviation from the lane center.
Checks a local CARLA checkout for missing API documentation, outdated references, broken links, and pages absent from navigation. Produces a grouped Markdown report without building the project or running the simulator.
Read and display CARLA object bounding boxes, then project them onto a camera image as 2D rectangles. Export an annotated PNG and box coordinates in JSON for object detection annotation work.
Builds CARLA's ue4-dev branch on Linux, producing the Unreal Engine 4.26 editor, server modules, and an installed Python client package. Fetches required content, optionally compiles native ROS 2 support, and verifies the build against a source-built server.