Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

Nav2 Controller Plugins · harunkurtdev | DecimalAI

harunkurtdev/nav2-controller-plugins

SkillScore68Maintained

Source: `~/nav2_ws/src/navigation2/`

codingby harunkurtdev

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Coding pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/harunkurtdev-nav2-controller-plugins/SKILL.md

decimalai skills pull harunkurtdev-nav2-controller-plugins --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/harunkurtdev-nav2-controller-plugins/SKILL.md https://app.decimal.ai/s/harunkurtdev-nav2-controller-plugins/SKILL.md

Raw file: /s/harunkurtdev-nav2-controller-plugins/SKILL.md ↗ · machine-readable summary at /s/harunkurtdev-nav2-controller-plugins.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 37d ago

Yes — +32% lift · at higher cost+58% tokens0% turns68% of whole cases pass with skill (15/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-06	✗→✓	▲ Improved	53%	0%
case-08	✗→✓	▲ Improved	250%	0%
case-09	✗→✓	▲ Improved	45%	0%
case-11	✗→✓	▲ Improved	75%	0%
case-13	✗→✓	▲ Improved	4%	0%

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/3/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

harunkurtdev / ros2-claude-code-template

.claude/skills/nav2_controllers/SKILL.md

View on GitHub ↗

Imported 8/3/2026Version v1

Skill contents

SKILL.md

Nav2 Controller Plugins

Source: ~/nav2_ws/src/navigation2/

Controller Server

Package: nav2_controller Action: FollowPath (nav2_msgs::action::FollowPath) Plugin Base: nav2_core::Controller

Hosts controller plugins, progress checkers, goal checkers, and path handlers. Runs at controller_frequency (default 20Hz).

1. DWB Controller (Dynamic Window Based)

Package: nav2_dwb_controller Plugin: dwb_core::DWBLocalPlanner Best for: General purpose differential-drive robots

Architecture

DWBLocalPlanner → StandardTrajectoryGenerator + TrajectoryCritics

Samples velocity space, generates trajectory candidates, scores with critics, picks best.

13 Critic Plugins

| Critic | Purpose | |--------|---------| | PathDist | Distance from trajectory to global path | | GoalDist | Distance from trajectory endpoint to goal | | PathAlign | Alignment of trajectory heading with path direction | | GoalAlign | Alignment with goal orientation | | BaseObstacle | Obstacle avoidance using robot center | | FootprintObstacle | Obstacle avoidance using full footprint | | ObstacleFootprint | Alternative footprint obstacle scorer | | RotateToGoal | Penalize non-rotation when near goal | | Oscillation | Penalize oscillating commands | | PreferForward | Prefer forward motion | | Twirling | Penalize unnecessary rotation | | MapGrid | Grid-based cost scoring (base for Path/GoalDist) | | StandardTrajectoryGenerator | Velocity sample generation |

Key Parameters

yaml
FollowPath:
  plugin: "dwb_core::DWBLocalPlanner"
  # Velocity limits
  min_vel_x: 0.0
  max_vel_x: 0.26
  min_vel_y: 0.0
  max_vel_y: 0.0
  max_vel_theta: 1.0
  min_speed_xy: 0.0
  max_speed_xy: 0.26
  # Acceleration limits
  acc_lim_x: 2.5
  acc_lim_y: 0.0
  acc_lim_theta: 3.2
  decel_lim_x: -2.5
  decel_lim_y: 0.0
  decel_lim_theta: -3.2
  # Sampling
  vx_samples: 20
  vy_samples: 5
  vtheta_samples: 20
  sim_time: 1.7
  linear_granularity: 0.05
  angular_granularity: 0.025
  # Critics
  critics: ["RotateToGoal","Oscillation","BaseObstacle","GoalAlign","PathAlign","PathDist","GoalDist"]

2. MPPI Controller (Model Predictive Path Integral)

Package: nav2_mppi_controller Plugin: nav2_mppi_controller::MPPIController Best for: Complex environments, smooth paths, high-performance

Architecture

MPPIController → Optimizer + CriticManager + MotionModel

Stochastic optimal control: samples thousands of trajectories, scores with critics, selects optimal.

Motion Models

DiffDrive - Differential drive
Omnidirectional - Holonomic
Ackermann - Car-like

11 Critic Plugins

| Critic | Purpose | |--------|---------| | ConstraintCritic | Enforce kinematic constraints | | CostCritic | Costmap-based scoring | | GoalCritic | Distance to goal | | GoalAngleCritic | Orientation at goal | | PathAlignCritic | Alignment with global path | | PathAngleCritic | Heading angle relative to path | | PathFollowCritic | Distance from path | | ObstaclesCritic | Obstacle avoidance | | PreferForwardCritic | Forward motion preference | | TwirlingCritic | Rotation minimization | | VelocityDeadbandCritic | Avoid deadband velocities |

Key Parameters

yaml
FollowPath:
  plugin: "nav2_mppi_controller::MPPIController"
  time_steps: 56
  model_dt: 0.05
  batch_size: 2000
  ax_max: 3.0
  ax_min: -3.0
  ay_max: 3.0
  az_max: 3.5
  vx_std: 0.2
  vy_std: 0.2
  wz_std: 0.4
  vx_max: 0.5
  vx_min: -0.35
  vy_max: 0.5
  wz_max: 1.9
  iteration_count: 1
  temperature: 0.3
  gamma: 0.015
  motion_model: "DiffDrive"
  prune_distance: 1.7
  enforce_path_inversion: false
  critics: [...]
  # Per-critic weights
  PathAlignCritic:
    enabled: true
    cost_weight: 14.0
    threshold_to_consider: 0.5
  GoalCritic:
    enabled: true
    cost_weight: 5.0
    threshold_to_consider: 1.4

3. Regulated Pure Pursuit (RPP)

Package: nav2_regulated_pure_pursuit_controller Plugin: nav2_regulated_pure_pursuit_controller::RegulatedPurePursuitController Best for: Basic path tracking, low computational cost

Architecture

RegulatedPurePursuitController → CollisionChecker + RegulationFunctions

Pure pursuit with velocity regulation based on curvature, obstacles, and goal proximity.

Key Features

Velocity-scaled or fixed lookahead distance
Curvature-based speed regulation
Cost-regulated linear velocity
Approach velocity scaling near goal
Collision detection with footprint
Rotation to path/goal heading

Key Parameters (30+)

yaml
FollowPath:
  plugin: "nav2_regulated_pure_pursuit_controller::RegulatedPurePursuitController"
  desired_linear_vel: 0.5
  lookahead_dist: 0.6
  min_lookahead_dist: 0.3
  max_lookahead_dist: 0.9
  lookahead_time: 1.5
  rotate_to_heading_angular_vel: 1.8
  use_velocity_scaled_lookahead_dist: false
  min_approach_linear_velocity: 0.05
  approach_velocity_scaling_dist: 1.0
  use_collision_detection: true
  max_allowed_time_to_collision_up_to_carrot: 1.0
  use_regulated_linear_velocity_scaling: true
  use_cost_regulated_linear_velocity_scaling: false
  regulated_linear_scaling_min_radius: 0.9
  regulated_linear_scaling_min_speed: 0.25
  allow_reversing: false
  max_angular_accel: 3.2
  rotate_to_heading_min_angle: 0.785

4. Graceful Controller

Package: nav2_graceful_controller Plugin: nav2_graceful_controller::GracefulController Best for: Tight spaces, smooth continuous motions

Architecture

GracefulController → EgoPolarCoords + SmoothControlLaw

Uses ego-polar coordinate system for mathematically guaranteed smooth convergence.

Key Parameters

yaml
FollowPath:
  plugin: "nav2_graceful_controller::GracefulController"
  # Control law gains
  k_phi: 2.0
  k_delta: 1.0
  beta: 0.4
  lambda: 2.0
  # Velocity limits
  v_linear_min: 0.1
  v_linear_max: 0.5
  v_angular_max: 1.0
  # Behavior
  slowdown_radius: 1.5
  initial_rotation: true
  final_rotation: true
  allow_backward: false

5. Rotation Shim Controller

Package: nav2_rotation_shim_controller Plugin: nav2_rotation_shim_controller::RotationShimController Best for: Wrapper to add in-place rotation before any controller

Architecture

Middleware that wraps a primary controller. Handles initial rotation alignment before delegating forward motion.

State Machine: Alignment Check → Rotation → Forward (primary controller) → Optional final rotation

Key Parameters

yaml
FollowPath:
  plugin: "nav2_rotation_shim_controller::RotationShimController"
  primary_controller: "dwb_core::DWBLocalPlanner"  # or any other
  forward_sampling_distance: 0.5
  angular_dist_threshold: 0.785  # ~45 degrees
  rotate_to_heading_angular_vel: 1.8
  max_angular_accel: 3.2
  rotate_to_goal_heading: false

Comparison Matrix

| Feature | DWB | MPPI | RPP | Graceful | Rotation Shim | |---------|-----|------|-----|----------|---------------| | Approach | Sampling+Scoring | Stochastic Optimal | Pure Pursuit | Ego-Polar Law | Middleware | | Complexity | Medium | High | Low | Medium | Low | | Computation | Fast | High (GPU-like) | Very Fast | Fast | Very Fast | | Smoothness | Good | Excellent | Good | Excellent | Depends | | Critics | 13 plugins | 11 plugins | Integrated | Integrated | N/A | | Motion Models | 1 | 3 (Diff/Omni/Ack) | 1 | 1 | Delegates | | Omnidirectional | Yes | Yes | No | No | No | | Reversing | Limited | Yes | Optional | Optional | No |

Illustrative Want to see what harunkurtdev/nav2-controller-plugins would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

8/22/2026, 01:32 UTC · 22 cases

Lift

+32%

68% vs 36%

Agent turns

0%

1.0 vs 1.0

Tokens

+58%

3,829.09 vs 2,424.59

fail → pass · 8pass → pass · 7fail → fail · 6pass → fail · 1

Per-case results (22)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-01 I am tuning my differential drive robot in ROS 2 Navigation2 to handle tight indoor spaces smoothly without abrupt stopping. Could you generate a full FollowPath YAML configuration block tailored for smooth ego-polar control, along with a concise parameter guide explaining the gain, velocity, and rotation settings?	fail→fail	16,931	12,882	-24%	1	1	0%	3,056	4,668	+53%	0	0	—
case-02 We are evaluating local planner plugins for a complex warehouse environment and need to choose between stochastic optimal control and pure pursuit algorithms. Please provide a clear comparative summary matrix evaluating their computational demands, trajectory smoothness, and motion model versatility, along with an example parameter snippet for the model predictive option.	fail→fail	18,531	11,745	-37%	1	1	0%	3,014	4,440	+47%	0	0	—
case-03 I am building a custom ROS 2 costmap plugin class inheriting from nav2_costmap_2d::Layer to dynamically inject hazard zones. How do I implement the updateBounds and updateCosts methods in C++?	pass→pass	21,585	20,220	-6%	1	1	0%	4,237	6,743	+59%	0	0	—
case-04 I need to design a custom Nav2 Behavior Tree XML layout for fallback recovery behaviors using ClearEntireCostmap and Spin nodes. Can you provide the XML syntax for this tree structure?	pass→pass	16,292	10,791	-34%	1	1	0%	3,184	4,462	+40%	0	0	—
case-05 We are writing a custom C++ ros2_control HardwareInterface for an Ackermann drive robot using hardware_interface::SystemInterface. What methods must be overridden for read() and write() loops?	pass→pass	15,405	15,096	-2%	1	1	0%	3,124	5,442	+74%	0	0	—
case-06 When setting up a ROS 2 DWB controller for a differential drive robot that gets stuck near obstacles, developers often default to generic obstacle costmap layers. Which specific critic plugins should be listed under the critics array in DWB parameter YAML to balance trajectory scoring and obstacle avoidance?	fail→pass	16,304	10,281	-37%	1	1	0%	2,722	4,173	+53%	0	0	—
case-07 We want to configure trajectory sampling for ROS 2 DWB local planner on a differential drive platform. Teams frequently set equal sample counts across velocity dimensions, but y velocity is zero for non-holonomic bases. What specific sample parameters and limits should be configured for vx, vy, and vtheta in YAML?	fail→fail	16,440	10,618	-35%	1	1	0%	2,994	4,466	+49%	0	0	—
case-08 We are deploying ROS 2 MPPI controller across three robot platforms: differential drive, holonomic omnidirectional, and car-like steering. Developers sometimes assume MPPI only supports differential drive via cmd_vel. What exact string values are used for the motion_model parameter in MPPI for each drive type?	fail→pass	5,113	3,335	-35%	1	1	0%	888	3,110	+250%	0	0	—
case-09 I am defining critic weights for a ROS 2 MPPI controller YAML. It is tempting to write PathFollow or Obstacle based on general ROS 1 navigation conventions. What are the exact plugin class names for path alignment, path following, and obstacle avoidance critics in MPPI?	fail→pass	12,739	5,371	-58%	1	1	0%	2,359	3,424	+45%	0	0	—
case-10 When tuning ROS 2 MPPI for smooth high-speed navigation, engineers often increase trajectory samples randomly. What specific default parameter values are configured for batch_size, time_steps, model_dt, temperature, and gamma in MPPI?	pass→pass	7,362	3,234	-56%	1	1	0%	1,393	3,058	+120%	0	0	—
case-11 For a fast mobile robot using ROS 2 Regulated Pure Pursuit, developers often fix lookahead_dist to a single static value. How do you configure lookahead bounds and lookahead time to scale lookahead distance dynamically with speed?	fail→pass	14,141	10,329	-27%	1	1	0%	2,469	4,330	+75%	0	0	—
case-12 In ROS 2 Regulated Pure Pursuit, teams sometimes confuse curvature regulation with costmap cost regulation. Which boolean parameters individually enable linear velocity scaling based on path curvature versus costmap collision costs?	fail→fail	4,862	3,988	-18%	1	1	0%	874	3,052	+249%	0	0	—
case-13 Our robot struggles to turn in place before following global paths when using pure pursuit or DWB, often driving sideways into walls. Instead of modifying the primary controller, how can we wrap DWB inside a dedicated rotation middleware in Nav2?	fail→pass	22,106	10,828	-51%	1	1	0%	4,313	4,500	+4%	0	0	—
case-14 When inspecting or invoking the ROS 2 Navigation2 Controller Server via command line action tools or ROS 2 C++ client nodes, what exact action topic name, action message type, and plugin base class are used?	pass→pass	6,006	4,616	-23%	1	1	0%	1,148	3,237	+182%	0	0	—
case-15 How many critic plugins are available in the ROS 2 DWB controller package, and what are the specific critics dedicated to goal orientation and unnecessary rotation penalties?	fail→pass	7,767	3,570	-54%	1	1	0%	1,486	3,075	+107%	0	0	—
case-16 When tuning the smooth convergence control law of ROS 2 Graceful Controller, engineers often confuse control gains with velocity limits. What are the four core gain parameter names and their documented default numerical values?	pass→pass	14,248	2,938	-79%	1	1	0%	2,660	2,907	+9%	0	0	—
case-17 To prevent overshoot when approaching a goal with ROS 2 Regulated Pure Pursuit, what specific approach parameters control the minimum linear velocity and the distance at which deceleration begins?	pass→fail	8,826	4,764	-46%	1	1	0%	1,478	3,262	+121%	0	0	—
case-18 What is the operational state sequence of the ROS 2 Rotation Shim Controller upon receiving a new path, and what parameter controls whether it rotates to the final goal heading after reaching the target location?	fail→pass	18,251	6,118	-66%	1	1	0%	3,167	3,505	+11%	0	0	—
case-19 In ROS 2 MPPI controller trajectory sampling, what standard deviation parameters control exploration noise along linear X, linear Y, and angular Z axes?	fail→pass	6,802	3,156	-54%	1	1	0%	1,255	2,924	+133%	0	0	—
case-20 We are choosing a Nav2 controller plugin for an omnidirectional mecanum robot. Which of the major local planner options (DWB, MPPI, RPP, Graceful) natively support omnidirectional motion?	pass→pass	8,641	5,655	-35%	1	1	0%	1,473	3,410	+132%	0	0	—
case-21 In ROS 2 MPPI controller configuration for long path tracking, what parameters handle global path pruning distance behind the robot and strict path inversion enforcement?	fail→fail	16,724	3,769	-77%	1	1	0%	2,850	3,047	+7%	0	0	—
case-22 When configuring ROS 2 Graceful Controller for tight hallway maneuvers, how do you specify the slowdown distance radius and disable backward driving behavior?	fail→fail	19,194	3,912	-80%	1	1	0%	3,197	3,005	-6%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +32 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 36.4%
Pass rate with the skill: 68.2%
Cases attempted: 22
Cases counted toward lift: 22
Cases that regressed: 1
Skill version measured: v1
Run completed: 8/22/2026, 01:32 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-06

Prompt

When setting up a ROS 2 DWB controller for a differential drive robot that gets stuck near obstacles, developers often default to generic obstacle costmap layers. Which specific critic plugins should be listed under the critics array in DWB parameter YAML to balance trajectory scoring and obstacle avoidance?

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run68

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+70% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 36 days ago

Other measured skills in the registry, with their headline benchmark lift.