doc(tool-parity): live test of agy 1.1.9 — real toolset, schedule limits, buy-before-build notes

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

.gitea/workflows/release.yml

@@ -20,7 +20,7 @@ jobs:
       - name: Detektera ändrade tools
         id: changed
         run: |
-          TOOLS="pixel-sprite-maker mesh-tool bitmap-font-maker sfx-maker"
+          TOOLS="pixel-sprite-maker mesh-tool bitmap-font-maker sfx-maker hitbox-tool"
           BEFORE="${{ github.event.before }}"
           CHANGED=""
           if [ "${{ github.event_name }}" = "workflow_dispatch" ] \
@@ -58,6 +58,7 @@ jobs:
               mesh-tool)          BIN=mesht ;;
               bitmap-font-maker)  BIN=fontc ;;
               sfx-maker)          BIN=sfxc ;;
+              hitbox-tool)        BIN=hitbox ;;
               *) echo "okänt tool $t"; exit 1 ;;
             esac
             echo "=== $t ($BIN) ==="
@@ -86,6 +87,7 @@ jobs:
               mesh-tool)          BIN=mesht ;;
               bitmap-font-maker)  BIN=fontc ;;
               sfx-maker)          BIN=sfxc ;;
+              hitbox-tool)        BIN=hitbox ;;
             esac
             TAG="$t-latest"
             BODY="$t (binär: $BIN) - rullande bygge från senaste master. Commit: ${{ github.sha }}. Arkitekturer: linux x64 + arm64 (Pi5)."

.vscode/tasks.json

@@ -49,6 +49,22 @@
       "group": "test",
       "problemMatcher": ["$go"]
     },
+    {
+      "label": "build hitbox-tool",
+      "type": "shell",
+      "command": "go build -trimpath -ldflags '-s -w' -o build/hitbox .",
+      "options": { "cwd": "${workspaceFolder}/hitbox-tool" },
+      "group": "build",
+      "problemMatcher": ["$go"]
+    },
+    {
+      "label": "test hitbox-tool",
+      "type": "shell",
+      "command": "go test ./...",
+      "options": { "cwd": "${workspaceFolder}/hitbox-tool" },
+      "group": "test",
+      "problemMatcher": ["$go"]
+    },
     {
       "label": "test pixel-sprite-maker",
       "type": "shell",
@@ -67,7 +83,7 @@
     },
     {
       "label": "build",
-      "dependsOn": ["build pixel-sprite-maker", "build mesh-tool", "build bitmap-font-maker", "build sfx-maker"],
+      "dependsOn": ["build pixel-sprite-maker", "build mesh-tool", "build bitmap-font-maker", "build sfx-maker", "build hitbox-tool"],
       "dependsOrder": "parallel",
       "group": { "kind": "build", "isDefault": true },
       "problemMatcher": []

README.md

@@ -12,10 +12,20 @@ agent understands the result without opening an image viewer.
 | [`mesh-tool/`](mesh-tool/) | `mesht` | Creates, inspects and edits 3D models (OBJ + STL). ASCII multi-view rendering + measurements (bbox, volume, watertightness) let an agent *see* a model, edit it (scale/rotate/mirror/merge/primitives) and verify the result. |
 | [`bitmap-font-maker/`](bitmap-font-maker/) | `fontc` | Turns `.font` text files (pixel glyph grids, proportional widths) into font atlases (PNG + JSON metrics) and renders text strings to PNG or the terminal. |
 | [`sfx-maker/`](sfx-maker/) | `sfxc` | Synthesizes retro game sound effects (sfxr-style) from `.sfx` text presets to 16-bit WAV: waves, envelope, pitch slides, vibrato, arpeggio, filters. Deterministic, with built-in presets (jump, coin, laser…). |
+| [`hitbox-tool/`](hitbox-tool/) | `hitbox` | Scans sprite sheet PNGs and writes per-frame collision boxes as JSON from the alpha channel. Understands the spritec sheet naming convention including upscaled sheets. |
 
 Each tool has its own folder, its own README with the full format/CLI
 reference, its own tests and its own dev branch (`dev/<tool>`).
 
+## Planned: agent-capability & homelab-admin tools
+
+[`doc/tool-parity.md`](doc/tool-parity.md) compares Gemini CLI's
+built-in tools with Claude Code's, and specs the CLI tools that close
+the gaps (`notifyr`, `giteactl`, `waitfor`, `cronr`, `fleet`,
+`envaudit`, `reghelper`, `pagepub`, `nbcell`, `wtreectl`, `fanout`) so
+any agent gets the same capabilities via `run_shell_command`. Build
+order and rationale live there and in [`doc/plan.md`](doc/plan.md).
+
 ## Building
 
 ```bash

doc/plan.md

@@ -37,6 +37,26 @@ master/main. Only tools whose folders changed get rebuilt.
 - New tools: bitmap-font maker, sound-effect generator (sfxr-style text
   presets), tiled-map (.tmx) writer.
 
+### Agent-capability tools (see [tool-parity.md](tool-parity.md))
+
+Close the gap between Gemini CLI and Claude Code, plus homelab admin
+tools grounded in infra-Doc history. Suggested order:
+
+1. `notifyr` — send/read on the existing Pi5 ntfy bus (≈ PushNotification).
+2. `giteactl` — Gitea Actions runs, zst CI logs, `wait`/`wait-quiet`
+   build serialization, releases.
+3. `waitfor` — block until a condition holds (≈ Monitor);
+   `cronr` — systemd-user-timer scheduling (≈ CronCreate/List/Delete).
+4. `fleet` — one-shot homelab health snapshot via the read-only
+   claude-docker wrapper.
+5. `envaudit` — compose ↔ `.env` key/inline-secret audit;
+   `reghelper` — registry catalog + prune *plans*;
+   `pagepub` — publish HTML/MD report to a URL (≈ Artifact);
+   `nbcell` — Jupyter cell editing (≈ NotebookEdit);
+   `wtreectl` — disposable git worktrees.
+6. `fanout` — parallel headless-agent orchestration (≈ Workflow); only
+   on concrete need.
+
 ## Open questions
 
 - Versioned releases (`vX.Y.Z` tags per tool) on top of the rolling

doc/tool-parity.md

@@ -0,0 +1,297 @@
+# Tool parity: Google's agent vs Claude Code
+
+Goal: let Google's agent CLI do everything Claude Code can, by building
+the missing capabilities as small CLI tools in this repo. The agent
+calls them through its shell tool, so every tool follows the house
+rules: **text in, verifiable artifacts out**, single static Go binary,
+self-explanatory output.
+
+**Which Google agent?** Legacy `gemini-cli` is auth-dead for personal
+accounts (verified again 2026-08-05, see infra-Doc
+`hosts/brasse-linux01.md`). The real target is **`agy` (Antigravity
+CLI)** — live-tested in section 2, and its toolset differs from the
+old Gemini CLI docs. Section 1's table is kept for reference since
+Gemini CLI still exists in API-key mode.
+
+A design rule that fell out of the live test: **a dedicated binary
+beats ad-hoc shell because of approval prefixes.** agy (like Claude
+Code) allowlists commands by prefix — `notifyr …` can be approved once
+and forever, while every hand-rolled `for i in $(seq …); do curl …`
+loop is a unique string that needs fresh human approval. Small stable
+CLIs are therefore not just convenience: they are what makes
+unattended agent operation possible at all.
+
+Sources: Gemini CLI tools reference (<https://geminicli.com/docs/reference/tools/>),
+Claude Code's toolset as of 2026-08, live probing of `agy` 1.1.9.
+
+## 1. Already at parity — nothing to build
+
+| Capability | Claude Code | Gemini CLI |
+|---|---|---|
+| Read/write/edit files | `Read` / `Write` / `Edit` | `read_file` / `write_file` / `replace` |
+| Find files / search text / list dirs | `Glob` / `Grep` | `glob` / `grep_search` / `list_directory`, plus `read_many_files` |
+| Shell, incl. background processes | `Bash` (+ background tasks) | `run_shell_command` (+ background processes) |
+| Web | `WebFetch` / `WebSearch` | `web_fetch` / `google_web_search` |
+| Ask the user a structured question | `AskUserQuestion` | `ask_user` |
+| Plan mode | `EnterPlanMode` / `ExitPlanMode` | `enter_plan_mode` / `exit_plan_mode` |
+| Skills / slash commands | `Skill` (`.claude/skills`) | `activate_skill` (`.gemini/skills`) |
+| Persistent memory | file-based memory dir | `save_memory` (simpler, but exists) |
+| Todo/task tracking | `TaskCreate`/`TaskUpdate`/… | `write_todos`, `tracker_*` (experimental) |
+| MCP servers + resources | MCP tools, `ListMcpResources`/`ReadMcpResource` | MCP tools, `list_mcp_resources`/`read_mcp_resource` |
+| Subagents | `Agent` (background, custom types) | subagents (experimental) — weaker, see `fanout` below |
+
+Not worth replicating (harness-internal to Claude Code, no value as a
+CLI): `ToolSearch`, `EndConversation`, `ReportFindings`,
+`ShareOnboardingGuide`, `DesignSync`, remote cloud execution.
+
+## 2. Live test 2026-08-05: `agy` (Antigravity CLI 1.1.9)
+
+Tested interactively in a tmux session (Google AI Pro account, model
+Gemini 3.6 Flash). Its 19 built-in tools, self-enumerated:
+
+`ask_permission`, `ask_question`, `define_subagent`, `generate_image`,
+`grep_search`, `invoke_subagent`, `list_dir`, `list_permissions`,
+`manage_subagents`, `manage_task`, `multi_replace_file_content`,
+`read_url_content`, `replace_file_content`, `run_command`, `schedule`,
+`search_web`, `send_message`, `view_file`, `write_to_file`.
+
+What this changes vs the old Gemini CLI picture:
+
+- **agy has real subagents** (`invoke_subagent`/`define_subagent`/
+  `manage_subagents` + `send_message`) and background-task management
+  (`manage_task`). → `fanout` demoted further; probably never needed.
+- **agy has `schedule`** — one-shot timer or cron expression that wakes
+  the agent with a prompt (same idea as Claude's `ScheduleWakeup`).
+  Confirmed limits, from its schema: it cannot run commands itself,
+  it is **in-memory and dies with the session**, and it cannot reach
+  the phone. → `cronr` (persistent systemd timers) and `notifyr` are
+  still needed; `schedule` complements them within a session.
+- **agy has `generate_image`** — a *reverse* gap: Claude Code has no
+  native image generation. Nothing to build; just worth knowing.
+- No MCP-resource tools, no memory tool and no glob in its toolset
+  (grep/list_dir cover finding files).
+
+Behavior tests run in a scratch arena:
+
+| Test | Result |
+|---|---|
+| Enumerate tools | Clean list of 19 (above) |
+| Edit a text file | Worked, auto-approved in trusted folder |
+| Edit a Jupyter cell, keep `.ipynb` valid | **Passed** — it wrote a `python3 -c` json script rather than text-replacing. Notebook stayed valid. Cost: a per-command approval each time → `nbcell` demoted to nice-to-have (stable prefix + no ad-hoc python). |
+| Wait for a file to appear | Worked via a hand-rolled `for … sleep 1` shell loop — a unique command string needing fresh approval. → exactly the `waitfor` case. |
+| Asked agy which CLI tools *it* wants for the homelab | Its list: ntfy client, `tea` (Gitea CLI), `skopeo`/`crane` (registry), `ofelia`/cron daemon, `ctop`-style fleet status — near-1:1 with section 4, and it independently made the approval-prefix argument. |
+
+**Buy before build:** agy's suggestions overlap with off-the-shelf
+tools. Evaluate first: `tea` (official Gitea CLI — but it does not read
+the Pi5's zst action logs and has no `wait-quiet`, which stay
+`giteactl`'s reason to exist, possibly as a thin layer *on top of*
+`tea`), `skopeo`/`crane` (cover most of `reghelper` — remaining value
+is prune *plans* and size summaries), `ctop` (interactive TUI, not
+agent-friendly output — `fleet` still wins for agents).
+
+## 3. Gaps → tools to build
+
+Ordered by expected value. Each becomes its own folder + binary +
+README, per repo convention.
+
+### 3.1 `cronr` — scheduled/recurring runs *(Claude: `CronCreate`/`CronList`/`CronDelete`, `/loop`)*
+
+agy's built-in `schedule` dies with the session (see section 2).
+`cronr` manages **systemd user timers** so an agent can create
+recurring or one-shot jobs that survive session exit and reboot
+(including "run this prompt every morning" via `agy -p …`).
+
+```
+cronr add nightly-ci-check --schedule "*-*-* 07:00" --cmd 'agy -p "check CI status and notify"'
+cronr add once-reboot-check --at "2026-08-06 03:00" --cmd '…'   # one-shot
+cronr list                    # name, schedule, next run, last result
+cronr logs nightly-ci-check   # journalctl for the unit
+cronr rm nightly-ci-check
+```
+
+Output prints the generated unit files so the result is verifiable.
+No daemon of its own — systemd does the running.
+
+### 3.2 `waitfor` — block until a condition holds *(Claude: `Monitor`)*
+
+Turns "poll every N seconds" into a single blocking tool call, so the
+agent doesn't burn turns polling.
+
+```
+waitfor --cmd "curl -sf https://gitea.brasse-pc.eu/api/healthz" --interval 30s --timeout 20m
+waitfor --cmd "ssh pi5 docker ps --format '{{.Names}}'" --matches 'gitea' --timeout 10m
+waitfor … --then 'notifyr send --msg "gitea is back up"'
+```
+
+Exit 0 = condition met, exit 3 = timeout; last output is printed either
+way. `--then` runs a command on success (composes with `notifyr`).
+
+### 3.3 `notifyr` — push notifications, send **and read** *(Claude: `PushNotification`)*
+
+The ntfy server **already runs on the Pi5** and is the house-wide
+notification bus (topics like `Info`, `pi5-server-fel`, `ci-fel` — see
+infra-Doc `services/observability.md`). `notifyr` is a thin client so
+every agent uses it the same way:
+
+```
+notifyr send --topic ci-fel --title "Build failed" --msg "agent-tools arm64 test: FAIL" --priority high
+notifyr read --topic pi5-server-fel --since 2h     # poll mode: what has alerted lately?
+```
+
+`read` (ntfy's `?poll=1&since=…`) is the underrated half: it lets an
+agent *check what the infra has been complaining about* before/after a
+change. Config (`~/.config/notifyr/config.json`): server URL + token.
+
+### 3.4 `pagepub` — publish an HTML/Markdown report to a URL *(Claude: `Artifact`)*
+
+Claude Code can publish reports as web pages; Gemini cannot. `pagepub`
+rsyncs a file to a static-file host on the Pi5 (nginx container behind
+NPM, e.g. `pages.brasse-pc.eu`) and prints the stable URL.
+
+```
+pagepub publish report.html --slug ci-report     → https://pages.brasse-pc.eu/ci-report/
+pagepub publish notes.md --slug pi5-audit        # .md rendered to HTML with built-in template
+pagepub list | rm <slug>
+```
+
+Requires the static host to exist first (small infra task; goes in
+infra-Doc + NPM proxy host via `npmctl`).
+
+### 3.5 `nbcell` — Jupyter notebook editing *(Claude: `NotebookEdit`)*
+
+`.ipynb` is JSON that's miserable to edit via `replace`. `nbcell`
+exposes cells as text:
+
+```
+nbcell list nb.ipynb                 # index, type, first line, exec count
+nbcell show nb.ipynb 3               # cell source (and outputs with --outputs)
+nbcell edit nb.ipynb 3 --from-file cell.py
+nbcell add nb.ipynb --at 4 --type code --from-file new.py
+nbcell rm nb.ipynb 7
+```
+
+### 3.6 `wtreectl` — disposable git worktrees *(Claude: worktree isolation for agents)*
+
+Claude Code can give each subagent an isolated git worktree. `wtreectl`
+does the same for any agent:
+
+```
+wtreectl new [--branch dev/foo]   # prints the new worktree path
+wtreectl list
+wtreectl clean                    # removes worktrees with no changes
+```
+
+Lets two agent sessions work in the same repo without trampling each
+other.
+
+### 3.7 `fanout` — parallel subagent orchestration *(Claude: `Workflow`, `Agent`)* — roadmap, not agreed
+
+Runs N prompts as parallel headless agent processes (`gemini -p` /
+`claude -p`) with a concurrency cap, collecting each result as JSON in
+an output dir. A poor man's `Workflow`:
+
+```
+fanout run jobs.json --max 3 --out results/
+```
+
+Heavier than the other tools and overlaps with agent-helm's territory —
+park until there's a concrete need.
+
+## 4. Suggested tools from infra history
+
+Grounded in what the agent has already been doing per infra-Doc
+(`maintenance-and-gaps.md`, `services/source-control-and-deploy.md`,
+`services/observability.md`, the per-service "operational quick-ref"
+blocks). These help **any** agent (Claude or Gemini) administer the
+fleet, and they respect the read-only sudo policy
+(`ssh/claude-sudo-policy.md`): everything below is read-or-notify;
+mutations still go through Björn's supervised tmux flow.
+
+### 4.1 `giteactl` — Gitea repos, Actions runs and CI logs
+
+The biggest recurring friction. Today: CI status is polled ad hoc, and
+logs for private repos are only readable as zst files under
+`/srv/storage1/gitea/actions_log/…` on the Pi5. Wraps the Gitea REST +
+Actions API:
+
+```
+giteactl runs <repo> [--limit 5]          # status, branch, duration
+giteactl log <repo> <run> [--job N]       # fetches + decompresses the zst log
+giteactl wait <repo> [--timeout 30m]      # block until latest run finishes; exit 0 = green
+giteactl wait-quiet [--max-active 1]      # block until ≤N heavy builds are running
+giteactl release <repo> [<tag>]           # rolling-release assets + checksums
+```
+
+`wait-quiet` encodes the hard-learned rule "serialize pushes — >2 heavy
+builds take the Pi5 down": `giteactl wait-quiet && git push`.
+Composes with `waitfor`/`notifyr`.
+
+### 4.2 `fleet` — one-shot health snapshot of the whole homelab
+
+Every service doc ends with the same hand-rolled loop over
+`ssh pi5-claude sudo claude-docker ps/inspect …`. `fleet` does that
+loop once, properly:
+
+```
+fleet status          # containers (state, image, restarts), disk/mergerfs fill %, failed systemd units
+fleet status --host brasse-linux01
+fleet checks          # Uptime Kuma monitor states + last ntfy alerts (via notifyr read)
+```
+
+Read-only by construction (claude-docker wrapper + sudo allowlist), so
+it needs no new permissions. Output is a stable text table an agent can
+diff between runs.
+
+### 4.3 `envaudit` — compose ↔ `.env` key auditor
+
+Grounded in a real incident (`${STORAGE1}` undefined in `/srv/.env` →
+bad mount, rollback) and in maintenance-and-gaps' inline-secret
+findings:
+
+```
+envaudit check /srv/dockge-staks --env /srv/.env
+  → UNDEFINED  ${STORAGE1}   used by media-stack/compose.yaml
+  → INLINE     LDAP_ADMIN_PASSWORD  hardcoded in openldap/compose.yaml (should live in .env)
+  → UNUSED     OLD_API_KEY   defined but referenced nowhere
+```
+
+Pure text analysis of compose files — safe to run anywhere, catches the
+two failure classes that have actually happened.
+
+### 4.4 `reghelper` — docker-registry catalog & hygiene
+
+The LAN registry (`192.168.0.19:5000`) has no UI, no auth and no
+cleanup story, and the backup plan explicitly wants it slimmed:
+
+```
+reghelper ls                      # catalog + tags + image sizes
+reghelper tags <image>
+reghelper prune-plan --keep 2     # prints the delete+GC commands (does NOT run them)
+```
+
+`prune-plan` deliberately only *prints* the mutation commands for the
+supervised tmux flow — same pattern as the sudo policy.
+
+### 4.5 Backup status reader — once Backrest/restic exists
+
+future-plans.md has the whole Backrest+restic design chosen but
+unbuilt. When it lands, a `fleet backups` subcommand (last snapshot age
+per source, repo size, last check result) closes the loop — an agent
+can then *verify* backups instead of trusting them. Not a separate
+tool; park under `fleet`.
+
+### Cross-reference
+
+`/home/brasse/repos/dify-agent-tools/` already has a specced-but-unbuilt
+set of FastAPI tools (file/image sorting, face recognition, ST-card
+export…) for the Dify platform. Different runtime (HTTP tools vs CLI
+binaries), same philosophy — don't duplicate those here.
+
+## 5. Suggested build order
+
+1. `notifyr` — smallest, everything else composes with it, ntfy already runs.
+2. `giteactl` — removes the biggest daily friction (CI logs + build serialization).
+3. `waitfor` + `cronr` — turns both agents into unattended operators.
+4. `fleet` — replaces the hand-rolled health loops in every runbook.
+5. `envaudit`, `reghelper`, `pagepub`, `nbcell`, `wtreectl` — as needed.
+6. `fanout` — only if a concrete multi-agent need shows up.

hitbox-tool/README.md

@@ -0,0 +1,66 @@
+# hitbox-tool (`hitbox`)
+
+Scans **sprite sheet PNGs** and writes per-frame **collision boxes as
+JSON**, computed from the alpha channel. Closes the loop with
+[`pixel-sprite-maker`](../pixel-sprite-maker/): render a sheet with
+`spritec`, scan it with `hitbox`, and your game gets both graphics and
+hitboxes without a human drawing rectangles. Go, zero dependencies,
+single static binary.
+
+## Build
+
+```bash
+# Arch/Garuda: sudo pacman -S go
+cd hitbox-tool
+go build -o build/hitbox .        # or the VS Code task "build hitbox-tool"
+go test ./...
+```
+
+## Usage
+
+```bash
+# spritec-named sheets need no flags — layout is parsed from the name,
+# and integer upscales are auto-detected (a _8x8_4x1 sheet that is
+# 256x64 px was rendered at --scale 8, so cells are 64x64):
+hitbox scan walk_8x8_4x1.png -o walk.hitbox.json
+
+hitbox scan boss.png --cell 32x32 -o boss.json   # explicit frame size
+hitbox scan portrait.png                          # whole image = one frame, JSON to stdout
+
+hitbox show walk_8x8_4x1.png --frame 2            # verify visually in the terminal
+
+# tuning:
+--threshold 128     # ignore faint pixels (alpha < 128)
+--shrink 1          # tighter, more forgiving hitboxes (n px per side)
+--pad 2             # generous hitboxes (e.g. pickups)
+```
+
+## Output
+
+```json
+{
+  "image": "walk_8x8_4x1.png",
+  "cellW": 8, "cellH": 8, "cols": 4, "rows": 1,
+  "alphaThreshold": 1,
+  "frames": [
+    { "index": 0, "col": 0, "row": 0, "empty": false,
+      "box": { "x": 1, "y": 0, "w": 5, "h": 8 } },
+    { "index": 1, "col": 1, "row": 0, "empty": true, "box": null }
+  ]
+}
+```
+
+- Boxes are **relative to each frame's top-left corner**.
+- Frame order is row-major (`index = row * cols + col`), matching
+  spritec sheets.
+- Cells with no solid pixels get `"empty": true`.
+
+In game code:
+
+```
+hit = px >= frameX + box.x && px < frameX + box.x + box.w
+   && py >= frameY + box.y && py < frameY + box.y + box.h
+```
+
+`hitbox show` draws each frame with `#` for solid pixels and `+` for the
+box outline, so an agent can verify the result without an image viewer.

hitbox-tool/examples/walk_8x8_4x1.hitbox.json

@@ -0,0 +1,58 @@
+{
+  "image": "examples/walk_8x8_4x1.png",
+  "cellW": 64,
+  "cellH": 64,
+  "cols": 4,
+  "rows": 1,
+  "alphaThreshold": 1,
+  "frames": [
+    {
+      "index": 0,
+      "col": 0,
+      "row": 0,
+      "empty": false,
+      "box": {
+        "x": 8,
+        "y": 0,
+        "w": 40,
+        "h": 64
+      }
+    },
+    {
+      "index": 1,
+      "col": 1,
+      "row": 0,
+      "empty": false,
+      "box": {
+        "x": 8,
+        "y": 0,
+        "w": 40,
+        "h": 64
+      }
+    },
+    {
+      "index": 2,
+      "col": 2,
+      "row": 0,
+      "empty": false,
+      "box": {
+        "x": 8,
+        "y": 0,
+        "w": 40,
+        "h": 64
+      }
+    },
+    {
+      "index": 3,
+      "col": 3,
+      "row": 0,
+      "empty": false,
+      "box": {
+        "x": 8,
+        "y": 0,
+        "w": 40,
+        "h": 64
+      }
+    }
+  ]
+}

hitbox-tool/go.mod

@@ -0,0 +1,3 @@
+module gitea.brasse-pc.eu/brasse/agent-tools/hitbox-tool
+
+go 1.24

hitbox-tool/hitbox/hitbox.go

@@ -0,0 +1,230 @@
+// Package hitbox computes per-frame collision boxes from sprite sheet
+// images by scanning the alpha channel.
+package hitbox
+
+import (
+	"encoding/json"
+	"fmt"
+	"image"
+	"image/png"
+	"io"
+	"os"
+	"path/filepath"
+	"regexp"
+	"strconv"
+)
+
+// Box is a rectangle relative to its frame's top-left corner.
+type Box struct {
+	X int `json:"x"`
+	Y int `json:"y"`
+	W int `json:"w"`
+	H int `json:"h"`
+}
+
+// Frame is the scan result for one sheet cell.
+type Frame struct {
+	Index int  `json:"index"`
+	Col   int  `json:"col"`
+	Row   int  `json:"row"`
+	Empty bool `json:"empty"`
+	Box   *Box `json:"box"` // nil when Empty
+}
+
+// Sheet is the full scan result; the JSON deliverable.
+type Sheet struct {
+	Image     string  `json:"image"`
+	CellW     int     `json:"cellW"`
+	CellH     int     `json:"cellH"`
+	Cols      int     `json:"cols"`
+	Rows      int     `json:"rows"`
+	Threshold int     `json:"alphaThreshold"`
+	Frames    []Frame `json:"frames"`
+}
+
+// layoutRe matches the spritec sheet naming convention:
+// <base>_<cellW>x<cellH>_<cols>x<rows>.<ext>
+var layoutRe = regexp.MustCompile(`_(\d+)x(\d+)_(\d+)x(\d+)\.[A-Za-z]+$`)
+
+// LayoutFromName extracts cell size and grid from a spritec-style file
+// name. ok is false when the name doesn't follow the convention.
+func LayoutFromName(path string) (cellW, cellH, cols, rows int, ok bool) {
+	m := layoutRe.FindStringSubmatch(filepath.Base(path))
+	if m == nil {
+		return 0, 0, 0, 0, false
+	}
+	cellW, _ = strconv.Atoi(m[1])
+	cellH, _ = strconv.Atoi(m[2])
+	cols, _ = strconv.Atoi(m[3])
+	rows, _ = strconv.Atoi(m[4])
+	return cellW, cellH, cols, rows, true
+}
+
+// InferScale detects integer-upscaled sheets: when the image is exactly
+// s times bigger than the name-declared layout (both axes, s >= 1), the
+// real cell size is cell*s. Returns 0 when the layout doesn't fit.
+func InferScale(imgW, imgH, cellW, cellH, cols, rows int) int {
+	baseW, baseH := cellW*cols, cellH*rows
+	if baseW <= 0 || baseH <= 0 || imgW%baseW != 0 || imgH%baseH != 0 {
+		return 0
+	}
+	s := imgW / baseW
+	if s < 1 || imgH/baseH != s {
+		return 0
+	}
+	return s
+}
+
+// LoadPNG reads a PNG image from disk.
+func LoadPNG(path string) (image.Image, error) {
+	f, err := os.Open(path)
+	if err != nil {
+		return nil, err
+	}
+	defer f.Close()
+	img, err := png.Decode(f)
+	if err != nil {
+		return nil, fmt.Errorf("%s: %w (only PNG is supported — sheets need an alpha channel)", path, err)
+	}
+	return img, nil
+}
+
+// Scan computes a tight box per frame: the smallest rectangle covering
+// every pixel with alpha >= threshold. shrink/pad (in pixels) contract or
+// expand each box afterwards, clamped to the cell.
+func Scan(img image.Image, name string, cellW, cellH, threshold, shrink, pad int) (*Sheet, error) {
+	b := img.Bounds()
+	if cellW <= 0 || cellH <= 0 {
+		cellW, cellH = b.Dx(), b.Dy() // whole image = one frame
+	}
+	if b.Dx()%cellW != 0 || b.Dy()%cellH != 0 {
+		return nil, fmt.Errorf("image is %dx%d which is not divisible by cell %dx%d",
+			b.Dx(), b.Dy(), cellW, cellH)
+	}
+	if threshold < 1 || threshold > 255 {
+		return nil, fmt.Errorf("alpha threshold %d out of range 1-255", threshold)
+	}
+	cols, rows := b.Dx()/cellW, b.Dy()/cellH
+	sheet := &Sheet{
+		Image: name, CellW: cellW, CellH: cellH,
+		Cols: cols, Rows: rows, Threshold: threshold,
+	}
+	for row := 0; row < rows; row++ {
+		for col := 0; col < cols; col++ {
+			fr := Frame{Index: row*cols + col, Col: col, Row: row}
+			minX, minY := cellW, cellH
+			maxX, maxY := -1, -1
+			for y := 0; y < cellH; y++ {
+				for x := 0; x < cellW; x++ {
+					_, _, _, a := img.At(b.Min.X+col*cellW+x, b.Min.Y+row*cellH+y).RGBA()
+					if int(a>>8) >= threshold {
+						if x < minX {
+							minX = x
+						}
+						if y < minY {
+							minY = y
+						}
+						if x > maxX {
+							maxX = x
+						}
+						if y > maxY {
+							maxY = y
+						}
+					}
+				}
+			}
+			if maxX < 0 {
+				fr.Empty = true
+			} else {
+				box := Box{X: minX, Y: minY, W: maxX - minX + 1, H: maxY - minY + 1}
+				box = adjust(box, shrink-pad, cellW, cellH)
+				fr.Box = &box
+			}
+			sheet.Frames = append(sheet.Frames, fr)
+		}
+	}
+	return sheet, nil
+}
+
+// adjust contracts the box by delta px on every side (negative delta
+// expands), clamped to the cell and to a minimum size of 1x1.
+func adjust(b Box, delta, cellW, cellH int) Box {
+	b.X += delta
+	b.Y += delta
+	b.W -= 2 * delta
+	b.H -= 2 * delta
+	if b.X < 0 {
+		b.W += b.X
+		b.X = 0
+	}
+	if b.Y < 0 {
+		b.H += b.Y
+		b.Y = 0
+	}
+	if b.X+b.W > cellW {
+		b.W = cellW - b.X
+	}
+	if b.Y+b.H > cellH {
+		b.H = cellH - b.Y
+	}
+	if b.W < 1 {
+		b.W = 1
+		if b.X > cellW-1 {
+			b.X = cellW - 1
+		}
+	}
+	if b.H < 1 {
+		b.H = 1
+		if b.Y > cellH-1 {
+			b.Y = cellH - 1
+		}
+	}
+	return b
+}
+
+// WriteJSON encodes the sheet as indented JSON.
+func (s *Sheet) WriteJSON(w io.Writer) error {
+	enc := json.NewEncoder(w)
+	enc.SetIndent("", "  ")
+	return enc.Encode(s)
+}
+
+// Ascii draws one frame with '#' for solid pixels and '+' for the box
+// outline, so results can be verified in a terminal.
+func Ascii(img image.Image, s *Sheet, index, threshold int) string {
+	if index < 0 || index >= len(s.Frames) {
+		return "(no such frame)\n"
+	}
+	fr := s.Frames[index]
+	b := img.Bounds()
+	out := make([]rune, 0, (s.CellW+1)*s.CellH)
+	onBoxEdge := func(x, y int) bool {
+		if fr.Box == nil {
+			return false
+		}
+		bx := fr.Box
+		inX := x >= bx.X && x < bx.X+bx.W
+		inY := y >= bx.Y && y < bx.Y+bx.H
+		edgeX := x == bx.X || x == bx.X+bx.W-1
+		edgeY := y == bx.Y || y == bx.Y+bx.H-1
+		return (inX && inY) && (edgeX || edgeY)
+	}
+	for y := 0; y < s.CellH; y++ {
+		for x := 0; x < s.CellW; x++ {
+			_, _, _, a := img.At(b.Min.X+fr.Col*s.CellW+x, b.Min.Y+fr.Row*s.CellH+y).RGBA()
+			solid := int(a>>8) >= threshold
+			switch {
+			case solid && onBoxEdge(x, y):
+				out = append(out, '#')
+			case solid:
+				out = append(out, '#')
+			case onBoxEdge(x, y):
+				out = append(out, '+')
+			default:
+				out = append(out, '.')
+			}
+		}
+		out = append(out, '\n')
+	}
+	return string(out)
+}

hitbox-tool/hitbox/hitbox_test.go

@@ -0,0 +1,137 @@
+package hitbox
+
+import (
+	"bytes"
+	"encoding/json"
+	"image"
+	"image/color"
+	"testing"
+)
+
+// sheet4 builds a 2x1 sheet of 4x4 cells: frame 0 has a 2x2 blob at
+// (1,1); frame 1 is empty.
+func sheet4() image.Image {
+	img := image.NewNRGBA(image.Rect(0, 0, 8, 4))
+	for y := 1; y <= 2; y++ {
+		for x := 1; x <= 2; x++ {
+			img.SetNRGBA(x, y, color.NRGBA{255, 0, 0, 255})
+		}
+	}
+	return img
+}
+
+func TestScanTightBox(t *testing.T) {
+	s, err := Scan(sheet4(), "test.png", 4, 4, 1, 0, 0)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if s.Cols != 2 || s.Rows != 1 || len(s.Frames) != 2 {
+		t.Fatalf("layout %dx%d frames=%d", s.Cols, s.Rows, len(s.Frames))
+	}
+	f0 := s.Frames[0]
+	if f0.Empty || f0.Box == nil {
+		t.Fatal("frame 0 should have a box")
+	}
+	if *f0.Box != (Box{X: 1, Y: 1, W: 2, H: 2}) {
+		t.Errorf("frame 0 box = %+v", *f0.Box)
+	}
+	f1 := s.Frames[1]
+	if !f1.Empty || f1.Box != nil {
+		t.Errorf("frame 1 should be empty, got %+v", f1)
+	}
+}
+
+func TestScanWholeImageDefault(t *testing.T) {
+	s, err := Scan(sheet4(), "x.png", 0, 0, 1, 0, 0)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(s.Frames) != 1 || s.CellW != 8 || s.CellH != 4 {
+		t.Errorf("whole-image scan: %+v", s)
+	}
+}
+
+func TestScanErrors(t *testing.T) {
+	if _, err := Scan(sheet4(), "x.png", 3, 4, 1, 0, 0); err == nil {
+		t.Error("non-divisible cell size should error")
+	}
+	if _, err := Scan(sheet4(), "x.png", 4, 4, 0, 0, 0); err == nil {
+		t.Error("threshold 0 should error")
+	}
+}
+
+func TestShrinkAndPad(t *testing.T) {
+	img := image.NewNRGBA(image.Rect(0, 0, 4, 4))
+	for y := 0; y < 4; y++ {
+		for x := 0; x < 4; x++ {
+			img.SetNRGBA(x, y, color.NRGBA{0, 0, 0, 255})
+		}
+	}
+	s, _ := Scan(img, "x.png", 4, 4, 1, 1, 0) // shrink 1
+	if *s.Frames[0].Box != (Box{X: 1, Y: 1, W: 2, H: 2}) {
+		t.Errorf("shrunk box = %+v", *s.Frames[0].Box)
+	}
+	s, _ = Scan(img, "x.png", 4, 4, 1, 0, 3) // pad clamps to cell
+	if *s.Frames[0].Box != (Box{X: 0, Y: 0, W: 4, H: 4}) {
+		t.Errorf("padded box = %+v", *s.Frames[0].Box)
+	}
+	s, _ = Scan(img, "x.png", 4, 4, 1, 10, 0) // over-shrink -> min 1x1
+	if s.Frames[0].Box.W < 1 || s.Frames[0].Box.H < 1 {
+		t.Errorf("over-shrunk box = %+v", *s.Frames[0].Box)
+	}
+}
+
+func TestThreshold(t *testing.T) {
+	img := image.NewNRGBA(image.Rect(0, 0, 2, 1))
+	img.SetNRGBA(0, 0, color.NRGBA{0, 0, 0, 100})
+	img.SetNRGBA(1, 0, color.NRGBA{0, 0, 0, 200})
+	s, _ := Scan(img, "x.png", 2, 1, 150, 0, 0)
+	if *s.Frames[0].Box != (Box{X: 1, Y: 0, W: 1, H: 1}) {
+		t.Errorf("threshold box = %+v", *s.Frames[0].Box)
+	}
+	s, _ = Scan(img, "x.png", 2, 1, 50, 0, 0)
+	if *s.Frames[0].Box != (Box{X: 0, Y: 0, W: 2, H: 1}) {
+		t.Errorf("low-threshold box = %+v", *s.Frames[0].Box)
+	}
+}
+
+func TestLayoutFromName(t *testing.T) {
+	w, h, c, r, ok := LayoutFromName("/tmp/walk_8x8_4x1.png")
+	if !ok || w != 8 || h != 8 || c != 4 || r != 1 {
+		t.Errorf("parsed %d %d %d %d ok=%v", w, h, c, r, ok)
+	}
+	if _, _, _, _, ok := LayoutFromName("plain.png"); ok {
+		t.Error("plain.png should not match")
+	}
+}
+
+func TestInferScale(t *testing.T) {
+	// walk_8x8_4x1.png rendered at scale 8 -> 256x64
+	if s := InferScale(256, 64, 8, 8, 4, 1); s != 8 {
+		t.Errorf("scale = %d, want 8", s)
+	}
+	if s := InferScale(32, 8, 8, 8, 4, 1); s != 1 {
+		t.Errorf("unscaled = %d, want 1", s)
+	}
+	if s := InferScale(250, 64, 8, 8, 4, 1); s != 0 {
+		t.Errorf("non-divisible = %d, want 0", s)
+	}
+	if s := InferScale(256, 32, 8, 8, 4, 1); s != 0 {
+		t.Errorf("axis mismatch = %d, want 0", s)
+	}
+}
+
+func TestJSONShape(t *testing.T) {
+	s, _ := Scan(sheet4(), "test.png", 4, 4, 1, 0, 0)
+	var buf bytes.Buffer
+	if err := s.WriteJSON(&buf); err != nil {
+		t.Fatal(err)
+	}
+	var back Sheet
+	if err := json.Unmarshal(buf.Bytes(), &back); err != nil {
+		t.Fatalf("invalid json: %v", err)
+	}
+	if back.Frames[0].Box.W != 2 {
+		t.Errorf("json roundtrip box = %+v", back.Frames[0].Box)
+	}
+}

hitbox-tool/main.go

@@ -0,0 +1,204 @@
+// hitbox scans sprite sheet PNGs and writes per-frame collision boxes
+// as JSON, using the alpha channel to find each frame's solid pixels.
+package main
+
+import (
+	"flag"
+	"fmt"
+	"os"
+	"strconv"
+	"strings"
+
+	"gitea.brasse-pc.eu/brasse/agent-tools/hitbox-tool/hitbox"
+)
+
+var version = "dev"
+
+const usage = `hitbox - collision box annotator for sprite sheets
+
+Usage:
+  hitbox scan <sheet.png> [flags]     compute per-frame boxes -> JSON
+  hitbox show <sheet.png> [flags]     draw frames + boxes in the terminal
+  hitbox version
+
+Scan flags:
+  --cell <WxH>       frame size, e.g. 8x8. Default: parsed from the
+                     spritec naming convention <name>_<W>x<H>_<C>x<R>.png;
+                     if neither is given the whole image is one frame.
+  --threshold <n>    alpha 1-255 that counts as solid (default 1)
+  --shrink <n>       contract every box by n px per side (forgiving hits)
+  --pad <n>          expand every box by n px per side
+  -o <path>          write JSON here (default: stdout)
+
+Show flags: --cell, --threshold, plus
+  --frame <n>        only this frame index (default: all)
+
+Boxes are relative to each frame's top-left corner. Frame indices are
+row-major (index = row * cols + col), matching spritec sheets.
+In game code: hit if (px,py) inside (frameX + box.x, frameY + box.y,
+box.w, box.h).
+`
+
+func main() {
+	if len(os.Args) < 2 {
+		fmt.Print(usage)
+		os.Exit(2)
+	}
+	switch os.Args[1] {
+	case "scan":
+		cmdScan(os.Args[2:])
+	case "show":
+		cmdShow(os.Args[2:])
+	case "version", "--version", "-v":
+		fmt.Println("hitbox", version)
+	case "help", "--help", "-h":
+		fmt.Print(usage)
+	default:
+		die("unknown command %q — run 'hitbox help'", os.Args[1])
+	}
+}
+
+// parseInterspersed lets flags appear before or after positional args.
+func parseInterspersed(fs *flag.FlagSet, args []string) {
+	var flags, pos []string
+	for i := 0; i < len(args); i++ {
+		a := args[i]
+		if len(a) > 1 && a[0] == '-' {
+			flags = append(flags, a)
+			name := strings.TrimLeft(a, "-")
+			if !strings.Contains(name, "=") {
+				f := fs.Lookup(name)
+				isBool := false
+				if f != nil {
+					if bv, ok := f.Value.(interface{ IsBoolFlag() bool }); ok && bv.IsBoolFlag() {
+						isBool = true
+					}
+				}
+				if !isBool && i+1 < len(args) {
+					i++
+					flags = append(flags, args[i])
+				}
+			}
+		} else {
+			pos = append(pos, a)
+		}
+	}
+	fs.Parse(append(flags, pos...))
+}
+
+// parseCell resolves the frame size from --cell, or from the spritec
+// naming convention (auto-detecting integer upscales: a sheet named
+// _8x8_4x1 that is 256x64 px was rendered at scale 8, so cells are 64x64).
+func parseCell(spec, path string, imgW, imgH int) (int, int, error) {
+	if spec != "" {
+		parts := strings.SplitN(strings.ToLower(spec), "x", 2)
+		if len(parts) == 2 {
+			w, err1 := strconv.Atoi(parts[0])
+			h, err2 := strconv.Atoi(parts[1])
+			if err1 == nil && err2 == nil && w > 0 && h > 0 {
+				return w, h, nil
+			}
+		}
+		return 0, 0, fmt.Errorf("--cell %q must look like 8x8", spec)
+	}
+	if w, h, cols, rows, ok := hitbox.LayoutFromName(path); ok {
+		if s := hitbox.InferScale(imgW, imgH, w, h, cols, rows); s > 1 {
+			fmt.Fprintf(os.Stderr, "hitbox: image is %dx the named layout — using %dx%d cells\n", s, w*s, h*s)
+			return w * s, h * s, nil
+		}
+		return w, h, nil
+	}
+	return 0, 0, nil // whole image = one frame
+}
+
+func cmdScan(args []string) {
+	fs := flag.NewFlagSet("scan", flag.ExitOnError)
+	cell := fs.String("cell", "", "frame size WxH")
+	threshold := fs.Int("threshold", 1, "solid alpha 1-255")
+	shrink := fs.Int("shrink", 0, "contract boxes n px per side")
+	pad := fs.Int("pad", 0, "expand boxes n px per side")
+	out := fs.String("o", "", "output JSON path (default stdout)")
+	parseInterspersed(fs, args)
+	if fs.NArg() != 1 {
+		die("scan takes exactly one PNG file")
+	}
+	path := fs.Arg(0)
+	img, err := hitbox.LoadPNG(path)
+	if err != nil {
+		die("%v", err)
+	}
+	b := img.Bounds()
+	cw, ch, err := parseCell(*cell, path, b.Dx(), b.Dy())
+	if err != nil {
+		die("%v", err)
+	}
+	sheet, err := hitbox.Scan(img, path, cw, ch, *threshold, *shrink, *pad)
+	if err != nil {
+		die("%v", err)
+	}
+	if *out == "" {
+		if err := sheet.WriteJSON(os.Stdout); err != nil {
+			die("%v", err)
+		}
+		return
+	}
+	f, err := os.Create(*out)
+	if err != nil {
+		die("%v", err)
+	}
+	defer f.Close()
+	if err := sheet.WriteJSON(f); err != nil {
+		die("%v", err)
+	}
+	solid := 0
+	for _, fr := range sheet.Frames {
+		if !fr.Empty {
+			solid++
+		}
+	}
+	fmt.Printf("%s  (%d frames of %dx%d, %d with pixels)\n",
+		*out, len(sheet.Frames), sheet.CellW, sheet.CellH, solid)
+}
+
+func cmdShow(args []string) {
+	fs := flag.NewFlagSet("show", flag.ExitOnError)
+	cell := fs.String("cell", "", "frame size WxH")
+	threshold := fs.Int("threshold", 1, "solid alpha 1-255")
+	frame := fs.Int("frame", -1, "frame index (default: all)")
+	parseInterspersed(fs, args)
+	if fs.NArg() != 1 {
+		die("show takes exactly one PNG file")
+	}
+	path := fs.Arg(0)
+	img, err := hitbox.LoadPNG(path)
+	if err != nil {
+		die("%v", err)
+	}
+	b := img.Bounds()
+	cw, ch, err := parseCell(*cell, path, b.Dx(), b.Dy())
+	if err != nil {
+		die("%v", err)
+	}
+	sheet, err := hitbox.Scan(img, path, cw, ch, *threshold, 0, 0)
+	if err != nil {
+		die("%v", err)
+	}
+	for _, fr := range sheet.Frames {
+		if *frame >= 0 && fr.Index != *frame {
+			continue
+		}
+		if fr.Empty {
+			fmt.Printf("frame %d (col %d, row %d): empty\n\n", fr.Index, fr.Col, fr.Row)
+			continue
+		}
+		fmt.Printf("frame %d (col %d, row %d): box x=%d y=%d w=%d h=%d\n",
+			fr.Index, fr.Col, fr.Row, fr.Box.X, fr.Box.Y, fr.Box.W, fr.Box.H)
+		fmt.Print(hitbox.Ascii(img, sheet, fr.Index, *threshold))
+		fmt.Println()
+	}
+}
+
+func die(format string, a ...any) {
+	fmt.Fprintf(os.Stderr, "hitbox: "+format+"\n", a...)
+	os.Exit(1)
+}