Update on 2.2.1

.claude-plugin/marketplace.json

@@ -12,7 +12,7 @@
       "name": "token-saver",
       "source": "./",
       "description": "Automatically compresses verbose CLI output to save tokens. 21 specialized processors for git, docker, npm, terraform, kubectl, helm, ansible, and more.",
-      "version": "2.1.1",
+      "version": "2.2.1",
       "author": {
         "name": "ppgranger"
       },

.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "token-saver",
   "description": "Automatically compresses verbose CLI output (git, docker, npm, terraform, kubectl, etc.) to save tokens in Claude Code sessions. 21 specialized processors with content-aware compression.",
-  "version": "2.1.1",
+  "version": "2.2.1",
   "author": {
     "name": "ppgranger",
     "url": "https://github.com/ppgranger"

README.md

@@ -237,24 +237,28 @@ processor is in [`docs/processors/`](docs/processors/).
 | 1 | **Package List** | 15 | pip list/freeze, npm ls, conda list, gem list, brew list | [package_list.md](docs/processors/package_list.md) |
 | 2 | **Git** | 20 | status, diff, log, show, push/pull/fetch, branch, stash, reflog, blame, cherry-pick, rebase, merge | [git.md](docs/processors/git.md) |
 | 3 | **Test** | 21 | pytest, jest, vitest, mocha, cargo test, go test, rspec, phpunit, bun test, npm/yarn/pnpm test, dotnet test, swift test, mix test | [test_output.md](docs/processors/test_output.md) |
-| 4 | **Build** | 25 | npm/yarn/pnpm build/install, cargo build, make, cmake, gradle, mvn, pip install, tsc, webpack, vite, next build, turbo, nx, bazel, sbt, mix compile, docker build | [build_output.md](docs/processors/build_output.md) |
-| 5 | **Lint** | 27 | eslint, ruff, flake8, pylint, clippy, mypy, prettier, biome, shellcheck, hadolint, rubocop, golangci-lint | [lint_output.md](docs/processors/lint_output.md) |
-| 6 | **Network** | 30 | curl, wget, http/https (httpie) | [network.md](docs/processors/network.md) |
-| 7 | **Docker** | 31 | ps, images, logs, pull/push, inspect, stats, compose up/down/build/ps/logs | [docker.md](docs/processors/docker.md) |
-| 8 | **Kubernetes** | 32 | kubectl/oc get, describe, logs, top, apply, delete, create | [kubectl.md](docs/processors/kubectl.md) |
-| 9 | **Terraform** | 33 | terraform/tofu plan, apply, destroy, init, output, state list/show | [terraform.md](docs/processors/terraform.md) |
-| 10 | **Environment** | 34 | env, printenv (with secret redaction) | [env.md](docs/processors/env.md) |
-| 11 | **Search** | 35 | grep -r, rg, ag, fd, fdfind | [search.md](docs/processors/search.md) |
-| 12 | **System Info** | 36 | du, wc, df | [system_info.md](docs/processors/system_info.md) |
-| 13 | **GitHub CLI** | 37 | gh pr/issue/run list/view/diff/checks/status | [gh.md](docs/processors/gh.md) |
-| 14 | **Database Query** | 38 | psql, mysql, sqlite3, pgcli, mycli, litecli | [db_query.md](docs/processors/db_query.md) |
-| 15 | **Cloud CLI** | 39 | aws, gcloud, az (JSON/table/text output compression) | [cloud_cli.md](docs/processors/cloud_cli.md) |
-| 16 | **Ansible** | 40 | ansible-playbook, ansible (ok/skipped counting, error preservation) | [ansible.md](docs/processors/ansible.md) |
-| 17 | **Helm** | 41 | helm install/upgrade/list/template/status/history | [helm.md](docs/processors/helm.md) |
-| 18 | **Syslog** | 42 | journalctl, dmesg (head/tail with error extraction) | [syslog.md](docs/processors/syslog.md) |
-| 19 | **File Listing** | 50 | ls, find, tree, exa, eza, rsync | [file_listing.md](docs/processors/file_listing.md) |
-| 20 | **File Content** | 51 | cat, head, tail, bat, less, more (content-aware: code, config, log, CSV) | [file_content.md](docs/processors/file_content.md) |
-| 21 | **Generic** | 999 | Any command (fallback: ANSI strip, dedup, truncation) | [generic.md](docs/processors/generic.md) |
+| 4 | **Python Install** | 24 | pip install, poetry install/update/add, uv pip install, uv sync | [python_install.md](docs/processors/python_install.md) |
+| 5 | **Build** | 25 | npm/yarn/pnpm build/install, cargo build, make, cmake, tsc, webpack, vite, next build, turbo, nx, bazel, sbt, mix compile, docker build | [build_output.md](docs/processors/build_output.md) |
+| 6 | **Cargo Clippy** | 26 | cargo clippy (multi-line block grouping with span/help preservation) | [cargo_clippy.md](docs/processors/cargo_clippy.md) |
+| 7 | **Lint** | 27 | eslint, ruff, flake8, pylint, clippy, mypy, prettier, biome, shellcheck, hadolint, rubocop, golangci-lint | [lint_output.md](docs/processors/lint_output.md) |
+| 8 | **Maven/Gradle** | 28 | mvn, ./mvnw, gradle, ./gradlew (download stripping, task noise removal) | [maven_gradle.md](docs/processors/maven_gradle.md) |
+| 9 | **Network** | 30 | curl, wget, http/https (httpie) | [network.md](docs/processors/network.md) |
+| 10 | **Docker** | 31 | ps, images, logs, pull/push, inspect, stats, compose up/down/build/ps/logs | [docker.md](docs/processors/docker.md) |
+| 11 | **Kubernetes** | 32 | kubectl/oc get, describe, logs, top, apply, delete, create | [kubectl.md](docs/processors/kubectl.md) |
+| 12 | **Terraform** | 33 | terraform/tofu plan, apply, destroy, init, output, state list/show | [terraform.md](docs/processors/terraform.md) |
+| 13 | **Environment** | 34 | env, printenv (with secret redaction) | [env.md](docs/processors/env.md) |
+| 14 | **Search** | 35 | grep -r, rg, ag, fd, fdfind | [search.md](docs/processors/search.md) |
+| 15 | **System Info** | 36 | du, wc, df | [system_info.md](docs/processors/system_info.md) |
+| 16 | **GitHub CLI** | 37 | gh pr/issue/run list/view/diff/checks/status | [gh.md](docs/processors/gh.md) |
+| 17 | **Database Query** | 38 | psql, mysql, sqlite3, pgcli, mycli, litecli | [db_query.md](docs/processors/db_query.md) |
+| 18 | **Cloud CLI** | 39 | aws, gcloud, az (JSON/table/text output compression) | [cloud_cli.md](docs/processors/cloud_cli.md) |
+| 19 | **Ansible** | 40 | ansible-playbook, ansible (ok/skipped counting, error preservation) | [ansible.md](docs/processors/ansible.md) |
+| 20 | **Helm** | 41 | helm install/upgrade/list/template/status/history | [helm.md](docs/processors/helm.md) |
+| 21 | **Syslog** | 42 | journalctl, dmesg (head/tail with error extraction) | [syslog.md](docs/processors/syslog.md) |
+| 22 | **Structured Log** | 45 | stern, kubetail (JSON Lines grouping by level) | [structured_log.md](docs/processors/structured_log.md) |
+| 23 | **File Listing** | 50 | ls, find, tree, exa, eza, rsync | [file_listing.md](docs/processors/file_listing.md) |
+| 24 | **File Content** | 51 | cat, head, tail, bat, less, more (content-aware: code, config, log, CSV) | [file_content.md](docs/processors/file_content.md) |
+| 25 | **Generic** | 999 | Any command (fallback: ANSI strip, dedup, truncation) | [generic.md](docs/processors/generic.md) |
 
 ## Configuration
 
@@ -341,11 +345,13 @@ Project settings are merged with global settings. Token-Saver walks up parent di
 | `max_traceback_lines` | 30 | Max traceback lines before truncation |
 | `db_prune_days` | 90 | Stats retention in days |
 | `user_processors_dir` | `~/.token-saver/processors/` | Directory for custom processors |
+| `disabled_processors` | `[]` | List of processor names to disable (env: comma-separated) |
+| `max_chain_depth` | 3 | Maximum processor chain depth |
 | `debug` | false | Enable debug logging |
 
 ## Custom Processors
 
-You can extend Token-Saver with your own processors for commands not covered by the built-in 21.
+You can extend Token-Saver with your own processors for commands not covered by the built-in 25.
 
 1. Create a Python file with a class inheriting from `src.processors.base.Processor`
 2. Implement `can_handle()`, `process()`, `name`, and set `priority`

docs/processors/cargo.md

@@ -0,0 +1,34 @@
+# Cargo Processor
+
+**File:** `src/processors/cargo.py` | **Priority:** 22 | **Name:** `cargo`
+
+Dedicated processor for Rust's cargo build system.
+
+## Supported Commands
+
+cargo build, cargo check, cargo doc, cargo update, cargo bench.
+
+## Strategy
+
+| Subcommand | Strategy |
+|---|---|
+| **build/check** | Collapse `Compiling X v1.0` lines into count. Group warnings by type (unused_variable, unused_import, dead_code, unused_mut, lifetime, borrow_checker). Show first N examples per type. Keep ALL errors with full span context (`-->`, `|`, `^^` markers). Keep `Finished` summary |
+| **doc** | Collapse `Documenting X` and `Compiling X` lines. Keep doc warnings, errors, `Finished`, and `Generated` lines |
+| **update** | Show all major version bumps explicitly (breaking changes). Collapse minor/patch bumps into count. Keep `Adding` and `Removing` lines |
+| **bench** | Keep benchmark result lines (`bench: N ns/iter`). Strip `Compiling` and `Running` noise. Keep `test result:` summary |
+
+## Exclusions
+
+- `cargo test` is routed to `TestOutputProcessor`
+- `cargo clippy` is routed to `CargoClippyProcessor`
+
+## Configuration
+
+| Parameter | Default | Description |
+|---|---|---|
+| cargo_warning_example_count | 2 | Number of example warnings to show per category |
+| cargo_warning_group_threshold | 3 | Minimum occurrences before warnings are grouped |
+
+## Removed Noise
+
+`Compiling X v1.0.0` lines, `Downloading X v1.0.0` lines, `Running` lines (bench), intermediate blank lines between warnings.

docs/processors/cargo_clippy.md

@@ -0,0 +1,39 @@
+# Cargo Clippy Processor
+
+**File:** `src/processors/cargo_clippy.py` | **Priority:** 26 | **Name:** `cargo_clippy`
+
+Dedicated processor for Rust clippy lint output with multi-line block awareness.
+
+## Supported Commands
+
+cargo clippy (with any flags like `--all-targets`, `-- -W clippy::all`).
+
+## Strategy
+
+Parses clippy's multi-line warning blocks (header + `-->` span + code + `= help:` annotations) as coherent units. Groups warnings by clippy lint rule. Shows N example blocks per rule with full context. Preserves all errors in full.
+
+| Output Type | Strategy |
+|---|---|
+| **Warnings** | Group by lint rule (e.g., `clippy::needless_return`). Show count + N example blocks per rule. Categorize as style/correctness/complexity/perf |
+| **Errors** | Keep all error blocks in full with spans and context |
+| **Checking/Compiling** | Collapse into count (e.g., `[12 checked, 3 compiled]`) |
+| **Summary** | Keep `warning: X generated N warnings` summary line |
+
+## Key Difference from Lint Processor
+
+The generic `LintOutputProcessor` groups violations as single lines. Clippy output has multi-line blocks with `-->` spans, code snippets, and `= help:` annotations that need to be preserved as coherent units. This processor keeps the block structure intact.
+
+## Configuration
+
+| Parameter | Default | Description |
+|---|---|---|
+| cargo_warning_example_count | 2 | Number of example warning blocks to show per rule |
+| cargo_warning_group_threshold | 3 | Minimum occurrences before warnings are grouped |
+
+## Chaining
+
+After clippy-specific processing, output is chained to the `lint` processor (`chain_to = ["lint"]`). This allows any non-clippy-specific warnings in the output to be grouped by the generic lint rule parser.
+
+## Fallback
+
+If this processor is disabled, `cargo clippy` falls back to the `LintOutputProcessor` which handles it at a line-by-line level.

docs/processors/go.md

@@ -0,0 +1,36 @@
+# Go Processor
+
+**File:** `src/processors/go.py` | **Priority:** 23 | **Name:** `go`
+
+Dedicated processor for Go toolchain commands.
+
+## Supported Commands
+
+go build, go vet, go mod tidy, go mod download, go generate, go install.
+
+## Strategy
+
+| Subcommand | Strategy |
+|---|---|
+| **build/install** | Keep all `file.go:line:col: message` errors. For multi-package builds with many `# package` headers, truncate to first 3. Pass through unchanged if no errors (successful builds produce no output) |
+| **vet** | Group warnings by type (printf, unreachable, shadow, unused, nil, loop). Show first N examples per type. Keep `# package` headers for context |
+| **mod tidy/download** | Collapse `go: downloading X v1.0` lines into count. Keep `go: added/upgraded/downgraded/removed` lines (important dependency changes) |
+| **generate** | Collapse `running` lines into count. Keep errors and generator output |
+
+## Exclusions
+
+- `go test` is routed to `TestOutputProcessor`
+- `golangci-lint` is routed to `LintOutputProcessor`
+
+## Configuration
+
+Uses existing parameters:
+
+| Parameter | Default | Description |
+|---|---|---|
+| lint_example_count | 2 | Examples per warning type (go vet) |
+| lint_group_threshold | 3 | Minimum occurrences before grouping |
+
+## Removed Noise
+
+`go: downloading X` lines, `# package` headers (when redundant), `running` lines from go generate.

docs/processors/jq_yq.md

@@ -0,0 +1,29 @@
+# JQ/YQ Processor
+
+**File:** `src/processors/jq_yq.py` | **Priority:** 44 | **Name:** `jq_yq`
+
+Compresses large JSON and YAML outputs from jq and yq.
+
+## Supported Commands
+
+jq, yq.
+
+## Strategy
+
+| Output Type | Strategy |
+|---|---|
+| **Small output** (< 50 lines) | Pass through unchanged |
+| **jq JSON** | Parse as JSON, compress with `compress_json_value()` (truncate arrays > 5 items, summarize deeply nested objects). Re-serialize with indent |
+| **jq streaming** (one JSON per line) | Detect repeated structure (same keys), show first 3 + count. Fallback to head/tail |
+| **yq YAML** | Count top-level keys and list items. Collapse large arrays (> 3 items at same indent) to count. Add structure summary header |
+
+## Configuration
+
+| Parameter | Default | Description |
+|---|---|---|
+| jq_passthrough_threshold | 50 | Lines below which output passes through unchanged |
+
+## Notes
+
+- No runtime dependencies: JSON parsing uses stdlib `json` module, YAML uses heuristic analysis (no PyYAML dependency)
+- Streaming jq output (one value per line) is detected and compressed separately from single-document output

docs/processors/maven_gradle.md

@@ -0,0 +1,40 @@
+# Maven/Gradle Processor
+
+**File:** `src/processors/maven_gradle.py` | **Priority:** 28 | **Name:** `maven_gradle`
+
+Dedicated processor for Maven and Gradle build output.
+
+## Supported Commands
+
+mvn, ./mvnw, gradle, ./gradlew (all subcommands).
+
+## Strategy
+
+### Maven
+
+| Output Type | Strategy |
+|---|---|
+| **Download lines** | Strip `[INFO] Downloading from` and `[INFO] Downloaded from` lines. Show count |
+| **Module lines** | Count `[INFO] Building module-name` lines |
+| **Errors** | Keep all `[ERROR]` and `[FATAL]` lines |
+| **Warnings** | Keep first 5 `[WARNING]` lines, summarize rest |
+| **Test results** | Keep `Tests run: N, Failures: N` lines |
+| **Reactor summary** | Keep reactor summary block |
+| **Build result** | Keep `BUILD SUCCESS`/`BUILD FAILURE` and timing |
+
+### Gradle
+
+| Output Type | Strategy |
+|---|---|
+| **Task lines** | Strip `UP-TO-DATE`, `NO-SOURCE`, `SKIPPED`, `FROM-CACHE` tasks. Keep executed tasks. Show counts |
+| **Errors** | Keep `FAILURE:` blocks, error details, `What went wrong` sections |
+| **Test results** | Keep test result summary lines |
+| **Build result** | Keep `BUILD SUCCESSFUL`/`BUILD FAILED` and actionable task summary |
+
+## Configuration
+
+No dedicated configuration keys. Uses default compression thresholds.
+
+## Removed Noise
+
+Maven: `[INFO] Downloading/Downloaded` lines, separator lines (`-----`), empty `[INFO]` lines. Gradle: `UP-TO-DATE`/`NO-SOURCE` task lines, progress indicators.

docs/processors/python_install.md

@@ -0,0 +1,29 @@
+# Python Install Processor
+
+**File:** `src/processors/python_install.py` | **Priority:** 24 | **Name:** `python_install`
+
+Dedicated processor for Python package installation output.
+
+## Supported Commands
+
+pip install, pip3 install, poetry install/update/add, uv pip install, uv sync.
+
+## Strategy
+
+| Tool | Strategy |
+|---|---|
+| **pip install** | Strip `Collecting` and `Downloading` lines. Remove progress bars. Count packages installed. Show `already satisfied` count. Preserve all errors and warnings. Show installed package summary (first 10 + count) |
+| **poetry install/update/add** | Strip `Resolving dependencies` progress. Count installed/updated/removed packages. Show package names with versions. Preserve errors |
+| **uv pip install/sync** | Strip download progress. Keep `Resolved N packages` and `Installed N packages` summaries. Preserve errors |
+
+## Exclusions
+
+- `pip list` and `pip freeze` are routed to `PackageListProcessor`
+
+## Configuration
+
+No dedicated configuration keys. Uses default compression thresholds.
+
+## Removed Noise
+
+`Collecting X>=1.0` lines, `Downloading X-1.0.whl` lines, pip progress bars, `Installing collected packages:` line, `Using cached` lines, `Resolving dependencies...` output from poetry.

docs/processors/ssh.md

@@ -0,0 +1,29 @@
+# SSH Processor
+
+**File:** `src/processors/ssh.py` | **Priority:** 43 | **Name:** `ssh`
+
+Handles non-interactive SSH and SCP command output.
+
+## Supported Commands
+
+- `ssh host 'command'` or `ssh host "command"` (non-interactive SSH with quoted remote command)
+- `ssh -o Option=value host 'command'` (with SSH options)
+- `scp` (all forms — always non-interactive)
+
+## Not Supported
+
+- `ssh host` (interactive SSH — no remote command) remains excluded from compression
+
+## Strategy
+
+| Command | Strategy |
+|---|---|
+| **SSH remote** | Apply log-style compression: keep first 10 + last 20 lines, preserve error lines with context in the middle section |
+| **SCP** | Collapse progress bar lines (containing `%` and transfer rates) to final status per file. Keep error lines (permission denied, connection refused, etc.) |
+
+## How It Works
+
+The SSH/SCP exclusion in `hook_pretool.py` was narrowed from a blanket `ssh|scp` exclusion to only exclude interactive SSH (no quoted command). This allows:
+- `ssh host 'ls -la'` — compressed (non-interactive)
+- `scp file host:/path` — compressed (always non-interactive)
+- `ssh host` — still excluded (interactive)

docs/processors/structured_log.md

@@ -0,0 +1,35 @@
+# Structured Log Processor
+
+**File:** `src/processors/structured_log.py` | **Priority:** 45 | **Name:** `structured_log`
+
+Processor for JSON Lines log output from log tailing tools.
+
+## Supported Commands
+
+stern, kubetail.
+
+## Strategy
+
+| Content Type | Strategy |
+|---|---|
+| **JSON Lines (>50% valid JSON)** | Parse each JSON object. Group entries by log level (error/warn/info/debug/trace). Show count per level. Extract and display error messages (up to 10). Detect level from common keys: `level`, `severity`, `log_level`, `lvl` |
+| **Non-JSON output** | Fall back to log compression (head/tail with error preservation) |
+
+## Level Detection
+
+Checks these JSON keys in order: `level`, `severity`, `log_level`, `loglevel`, `lvl`, `log.level`. Falls back to regex matching on message content for `ERROR`/`WARN` patterns.
+
+## Message Extraction
+
+Checks these JSON keys in order: `msg`, `message`, `text`, `log`, `body`. Truncates messages longer than 200 characters.
+
+## Configuration
+
+| Parameter | Default | Description |
+|---|---|---|
+| kubectl_keep_head | 5 | Lines to keep from start (non-JSON fallback) |
+| kubectl_keep_tail | 10 | Lines to keep from end (non-JSON fallback) |
+
+## Future Use
+
+This processor can be activated via `chain_to` from other processors for outputs that contain embedded JSON Lines.

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "token-saver"
-version = "2.1.1"
+version = "2.2.1"
 requires-python = ">=3.10"
 
 [project.optional-dependencies]

scripts/hook_pretool.py

@@ -86,7 +86,7 @@ def _load_compressible_patterns() -> list[str]:
 EXCLUDED_PATTERNS = [
     r"(?<!['\"])\|(?!['\"])",  # unquoted pipe (complex pipelines)
     r"^\s*(vi|vim|nano|emacs|code)\b",
-    r"^\s*(ssh|scp)\b",
+    r"^\s*ssh\s+(?:-\S+\s+)*\S+\s*$",  # interactive ssh only (no remote command)
     r"^\s*rsync\b.*\S+:\S+",  # only exclude remote rsync (host:path)
     r"(?:^|\s)token[-_]saver\s",  # avoid wrapping token-saver CLI itself
     r"wrap\.py",
@@ -106,7 +106,7 @@ def _load_compressible_patterns() -> list[str]:
     r"<\(",  # process substitution
     r"^\s*sudo\b",
     r"^\s*(vi|vim|nano|emacs|code)\b",
-    r"^\s*(ssh|scp)\b",
+    r"^\s*ssh\s+(?:-\S+\s+)*\S+\s*$",  # interactive ssh only
     r"^\s*rsync\b.*\S+:\S+",  # only exclude remote rsync (host:path)
     r"^\s*env\s+\S+=",
     r"(?:^|\s)token[-_]saver\s",

src/__init__.py

@@ -1,6 +1,6 @@
 import os
 
-__version__ = "2.1.1"
+__version__ = "2.2.1"
 
 
 def data_dir() -> str: