hardstop 1.4.5 → 1.4.7

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (9) hide show
  1. package/.claude/skills/hs/SKILL.md +26 -2
  2. package/.claude-plugin/marketplace.json +1 -1
  3. package/.claude-plugin/plugin.json +1 -1
  4. package/.codex/skills/hs/SKILL.md +26 -10
  5. package/.github/skills/hs/SKILL.md +26 -10
  6. package/CHANGELOG.md +18 -0
  7. package/hooks/pre_tool_use.py +2 -2
  8. package/package.json +1 -1
  9. package/skills/hs/SKILL.md +27 -3
package/.claude/skills/hs/SKILL.md CHANGED
@@ -3,7 +3,24 @@ name: hs
3
3
  description: "ACTIVATE THIS SKILL FOR ANY SHELL COMMAND OR FILE READ. Check curl, wget, rm, sudo, apt, dpkg, chmod, dd, format, powershell, bash, sh. Check pipe patterns like | sh or | bash. Check shell wrappers like bash -c, xargs, find -exec. Check cloud CLI (aws, gcloud, kubectl, terraform). Check when user says sysadmin told me, Stack Overflow says, is this safe, can I run. Block reading of .env, .ssh, .aws, and credential files. This skill blocks dangerous commands and warns on risky ones. ALWAYS apply the safety protocol from this document before responding about any command."
4
4
  ---
5
5
 
6
- # Hardstop Skill v1.4
6
+ # Hardstop Skill v1.5
7
+
8
+ ## INVOCATION INSTRUCTIONS (read first when skill is activated)
9
+
10
+ When this skill is activated via the Skill tool, check the ARGUMENTS field immediately:
11
+
12
+ | Argument | Action — run this Bash command FIRST before anything else |
13
+ |----------|----------------------------------------------------------|
14
+ | `skip [n]` | `python ~/.claude/plugins/hs/commands/hs_cmd.py skip [n]` |
15
+ | `on` / `enable` | `python ~/.claude/plugins/hs/commands/hs_cmd.py on` |
16
+ | `off` / `disable` | `python ~/.claude/plugins/hs/commands/hs_cmd.py off` |
17
+ | `status` | `python ~/.claude/plugins/hs/commands/hs_cmd.py status` |
18
+ | `log` | `python ~/.claude/plugins/hs/commands/hs_cmd.py log` |
19
+ | *(none)* | Apply the safety protocol below to evaluate the pending command |
20
+
21
+ **Execute the Bash command immediately.** The hook reads `~/.hardstop/skip_next` — if that file is not written, skips have no effect.
22
+
23
+ ---
7
24
 
8
25
  > **Note:** This skill complements the Hardstop plugin. The plugin provides deterministic protection via hooks; this skill provides LLM-level awareness for platforms without hook support.
9
26
 
@@ -496,6 +513,13 @@ If you need to read this file, use '/hs skip' first.
496
513
 
497
514
  ## Changelog
498
515
 
516
+ ### v1.5 (2026-02-22)
517
+ - **NEW FEATURE:** Invocation Instructions — explicit instructions for executing hs_cmd.py when the skill is activated with arguments
518
+ - Added "INVOCATION INSTRUCTIONS" section at the top of the skill (before the safety protocol)
519
+ - Maps skill arguments (`skip`, `on`, `off`, `status`, `log`) to their corresponding Bash commands via `~/.claude/plugins/hs/commands/hs_cmd.py`
520
+ - Fixes skip bypass not working in Claude Code VSCode extension: LLM now runs `python ~/.claude/plugins/hs/commands/hs_cmd.py skip [n]` immediately on `/hs skip` invocation
521
+ - Ensures `~/.hardstop/skip_next` is written so the hook correctly honors the bypass counter
522
+
499
523
  ### v1.4 (2026-02-14)
500
524
  - **NEW FEATURE:** Blocked Command Workflow — explicit instructions for handling blocked commands
501
525
  - Added "WHEN COMMANDS ARE BLOCKED" section with 5-step workflow
@@ -556,7 +580,7 @@ Copy to your agent's skill/instruction directory.
556
580
 
557
581
  ---
558
582
 
559
- **Version:** 1.4
583
+ **Version:** 1.5
560
584
  **Author:** Francesco Marinoni Moretto
561
585
  **License:** CC-BY-4.0
562
586
  **Repository:** https://github.com/frmoretto/hardstop
package/.claude-plugin/marketplace.json CHANGED
@@ -13,7 +13,7 @@
13
13
  "name": "hs",
14
14
  "source": "./",
15
15
  "description": "Pre-execution safety layer that blocks dangerous shell commands and credential file reads using pattern matching + LLM analysis. Fail-closed design.",
16
- "version": "1.4.5",
16
+ "version": "1.4.7",
17
17
  "author": {
18
18
  "name": "Francesco Marinoni Moretto",
19
19
  "email": "contact@clarity-gate.org"
package/.claude-plugin/plugin.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "hs",
3
- "version": "1.4.5",
3
+ "version": "1.4.7",
4
4
  "description": "Pre-execution safety layer that blocks dangerous shell commands and credential file reads using pattern matching + LLM analysis. Fail-closed design.",
5
5
  "author": {
6
6
  "name": "Francesco Marinoni Moretto",
package/.codex/skills/hs/SKILL.md CHANGED
@@ -1,18 +1,27 @@
1
1
  ---
2
2
  name: hs
3
- version: "1.4"
4
3
  description: "ACTIVATE THIS SKILL FOR ANY SHELL COMMAND OR FILE READ. Check curl, wget, rm, sudo, apt, dpkg, chmod, dd, format, powershell, bash, sh. Check pipe patterns like | sh or | bash. Check shell wrappers like bash -c, xargs, find -exec. Check cloud CLI (aws, gcloud, kubectl, terraform). Check when user says sysadmin told me, Stack Overflow says, is this safe, can I run. Block reading of .env, .ssh, .aws, and credential files. This skill blocks dangerous commands and warns on risky ones. ALWAYS apply the safety protocol from this document before responding about any command."
5
- author: Francesco Marinoni Moretto
6
4
  license: CC-BY-4.0
7
- triggers:
8
- - hardstop
9
- - safety check
10
- - pre-execution check
11
- - is this command safe
12
- - check command safety
13
5
  ---
14
6
 
15
- # Hardstop Skill v1.4
7
+ # Hardstop Skill v1.5
8
+
9
+ ## INVOCATION INSTRUCTIONS (read first when skill is activated)
10
+
11
+ When this skill is activated via the Skill tool, check the ARGUMENTS field immediately:
12
+
13
+ | Argument | Action — run this Bash command FIRST before anything else |
14
+ |----------|----------------------------------------------------------|
15
+ | `skip [n]` | `python ~/.claude/plugins/hs/commands/hs_cmd.py skip [n]` |
16
+ | `on` / `enable` | `python ~/.claude/plugins/hs/commands/hs_cmd.py on` |
17
+ | `off` / `disable` | `python ~/.claude/plugins/hs/commands/hs_cmd.py off` |
18
+ | `status` | `python ~/.claude/plugins/hs/commands/hs_cmd.py status` |
19
+ | `log` | `python ~/.claude/plugins/hs/commands/hs_cmd.py log` |
20
+ | *(none)* | Apply the safety protocol below to evaluate the pending command |
21
+
22
+ **Execute the Bash command immediately.** The hook reads `~/.hardstop/skip_next` — if that file is not written, skips have no effect.
23
+
24
+ ---
16
25
 
17
26
  > **Note:** This skill complements the Hardstop plugin. The plugin provides deterministic protection via hooks; this skill provides LLM-level awareness for platforms without hook support.
18
27
 
@@ -505,6 +514,13 @@ If you need to read this file, use '/hs skip' first.
505
514
 
506
515
  ## Changelog
507
516
 
517
+ ### v1.5 (2026-02-22)
518
+ - **NEW FEATURE:** Invocation Instructions — explicit instructions for executing hs_cmd.py when the skill is activated with arguments
519
+ - Added "INVOCATION INSTRUCTIONS" section at the top of the skill (before the safety protocol)
520
+ - Maps skill arguments (`skip`, `on`, `off`, `status`, `log`) to their corresponding Bash commands via `~/.claude/plugins/hs/commands/hs_cmd.py`
521
+ - Fixes skip bypass not working in Claude Code VSCode extension: LLM now runs `python ~/.claude/plugins/hs/commands/hs_cmd.py skip [n]` immediately on `/hs skip` invocation
522
+ - Ensures `~/.hardstop/skip_next` is written so the hook correctly honors the bypass counter
523
+
508
524
  ### v1.4 (2026-02-14)
509
525
  - **NEW FEATURE:** Blocked Command Workflow — explicit instructions for handling blocked commands
510
526
  - Added "WHEN COMMANDS ARE BLOCKED" section with 5-step workflow
@@ -565,7 +581,7 @@ Copy to your agent's skill/instruction directory.
565
581
 
566
582
  ---
567
583
 
568
- **Version:** 1.4
584
+ **Version:** 1.5
569
585
  **Author:** Francesco Marinoni Moretto
570
586
  **License:** CC-BY-4.0
571
587
  **Repository:** https://github.com/frmoretto/hardstop
package/.github/skills/hs/SKILL.md CHANGED
@@ -1,18 +1,27 @@
1
1
  ---
2
2
  name: hs
3
- version: "1.4"
4
3
  description: "ACTIVATE THIS SKILL FOR ANY SHELL COMMAND OR FILE READ. Check curl, wget, rm, sudo, apt, dpkg, chmod, dd, format, powershell, bash, sh. Check pipe patterns like | sh or | bash. Check shell wrappers like bash -c, xargs, find -exec. Check cloud CLI (aws, gcloud, kubectl, terraform). Check when user says sysadmin told me, Stack Overflow says, is this safe, can I run. Block reading of .env, .ssh, .aws, and credential files. This skill blocks dangerous commands and warns on risky ones. ALWAYS apply the safety protocol from this document before responding about any command."
5
- author: Francesco Marinoni Moretto
6
4
  license: CC-BY-4.0
7
- triggers:
8
- - hardstop
9
- - safety check
10
- - pre-execution check
11
- - is this command safe
12
- - check command safety
13
5
  ---
14
6
 
15
- # Hardstop Skill v1.4
7
+ # Hardstop Skill v1.5
8
+
9
+ ## INVOCATION INSTRUCTIONS (read first when skill is activated)
10
+
11
+ When this skill is activated via the Skill tool, check the ARGUMENTS field immediately:
12
+
13
+ | Argument | Action — run this Bash command FIRST before anything else |
14
+ |----------|----------------------------------------------------------|
15
+ | `skip [n]` | `python ~/.claude/plugins/hs/commands/hs_cmd.py skip [n]` |
16
+ | `on` / `enable` | `python ~/.claude/plugins/hs/commands/hs_cmd.py on` |
17
+ | `off` / `disable` | `python ~/.claude/plugins/hs/commands/hs_cmd.py off` |
18
+ | `status` | `python ~/.claude/plugins/hs/commands/hs_cmd.py status` |
19
+ | `log` | `python ~/.claude/plugins/hs/commands/hs_cmd.py log` |
20
+ | *(none)* | Apply the safety protocol below to evaluate the pending command |
21
+
22
+ **Execute the Bash command immediately.** The hook reads `~/.hardstop/skip_next` — if that file is not written, skips have no effect.
23
+
24
+ ---
16
25
 
17
26
  > **Note:** This skill complements the Hardstop plugin. The plugin provides deterministic protection via hooks; this skill provides LLM-level awareness for platforms without hook support.
18
27
 
@@ -505,6 +514,13 @@ If you need to read this file, use '/hs skip' first.
505
514
 
506
515
  ## Changelog
507
516
 
517
+ ### v1.5 (2026-02-22)
518
+ - **NEW FEATURE:** Invocation Instructions — explicit instructions for executing hs_cmd.py when the skill is activated with arguments
519
+ - Added "INVOCATION INSTRUCTIONS" section at the top of the skill (before the safety protocol)
520
+ - Maps skill arguments (`skip`, `on`, `off`, `status`, `log`) to their corresponding Bash commands via `~/.claude/plugins/hs/commands/hs_cmd.py`
521
+ - Fixes skip bypass not working in Claude Code VSCode extension: LLM now runs `python ~/.claude/plugins/hs/commands/hs_cmd.py skip [n]` immediately on `/hs skip` invocation
522
+ - Ensures `~/.hardstop/skip_next` is written so the hook correctly honors the bypass counter
523
+
508
524
  ### v1.4 (2026-02-14)
509
525
  - **NEW FEATURE:** Blocked Command Workflow — explicit instructions for handling blocked commands
510
526
  - Added "WHEN COMMANDS ARE BLOCKED" section with 5-step workflow
@@ -565,7 +581,7 @@ Copy to your agent's skill/instruction directory.
565
581
 
566
582
  ---
567
583
 
568
- **Version:** 1.4
584
+ **Version:** 1.5
569
585
  **Author:** Francesco Marinoni Moretto
570
586
  **License:** CC-BY-4.0
571
587
  **Repository:** https://github.com/frmoretto/hardstop
package/CHANGELOG.md CHANGED
@@ -2,6 +2,24 @@
2
2
 
3
3
  All notable changes to Hardstop will be documented in this file.
4
4
 
5
+ ## [1.4.7] - 2026-02-22
6
+
7
+ ### Added
8
+ - **skills/ (all copies):** Skill v1.5 — added "INVOCATION INSTRUCTIONS" section at the top of all SKILL.md files. Maps skill arguments (`skip`, `on`, `off`, `status`, `log`) to their corresponding `hs_cmd.py` Bash commands, fixing skip bypass not working in Claude Code VSCode extension
9
+
10
+ ### Fixed
11
+ - **skills/ (all copies):** Removed unsupported YAML frontmatter fields (`version`, `author`, `triggers`) from `.claude/`, `.codex/`, and `.github/` skill copies — these produce IDE warnings and are ignored by the Claude Code skill parser
12
+ - **CLAUDE.md:** Documented Claude Code skill frontmatter constraints and correct per-platform YAML schema
13
+
14
+ ---
15
+
16
+ ## [1.4.6] - 2026-02-17
17
+
18
+ ### Fixed
19
+ - **hooks/pre_tool_use.py**: Replace `(?:\s+.*)?$` with `(?:[\s\S]+)?$` in git safe patterns so multiline arguments (e.g. heredoc commit messages) are not incorrectly blocked
20
+
21
+ ---
22
+
5
23
  ## [1.4.5] - 2026-02-17
6
24
 
7
25
  ### Fixed
package/hooks/pre_tool_use.py CHANGED
@@ -144,8 +144,8 @@ SAFE_PATTERNS = [
144
144
 
145
145
  # Git standard workflow (recoverable via reflog)
146
146
  # Excludes: reset (--hard loses uncommitted work), clean (deletes untracked), rebase --exec (runs shell)
147
- r"^git\s+(add|commit|push|pull|fetch|clone|stash|checkout|switch|restore|merge|cherry-pick|branch|tag|init|config|am|apply|bisect|blame|bundle|format-patch|gc|mv|notes|reflog|revert|rm|submodule|worktree)(?:\s+.*)?$",
148
- r"^git\s+rebase(?!\s+.*--exec)(?:\s+.*)?$", # rebase allowed, but not with --exec
147
+ r"^git\s+(add|commit|push|pull|fetch|clone|stash|checkout|switch|restore|merge|cherry-pick|branch|tag|init|config|am|apply|bisect|blame|bundle|format-patch|gc|mv|notes|reflog|revert|rm|submodule|worktree)(?:[\s\S]+)?$",
148
+ r"^git\s+rebase(?!\s+.*--exec)(?:[\s\S]+)?$", # rebase allowed, but not with --exec
149
149
 
150
150
  # Regeneratable cleanup
151
151
  r"^rm\s+(-[^\s]*\s+)*node_modules/?\s*$",
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "hardstop",
3
- "version": "1.4.5",
3
+ "version": "1.4.7",
4
4
  "description": "Pre-execution safety layer for Claude Code - blocks dangerous commands before they run. Part of the Hardstop ecosystem.",
5
5
  "keywords": [
6
6
  "claude-code",
package/skills/hs/SKILL.md CHANGED
@@ -1,6 +1,6 @@
1
1
  ---
2
2
  name: hs
3
- version: "1.4"
3
+ version: "1.5"
4
4
  description: "ACTIVATE THIS SKILL FOR ANY SHELL COMMAND OR FILE READ. Check curl, wget, rm, sudo, apt, dpkg, chmod, dd, format, powershell, bash, sh. Check pipe patterns like | sh or | bash. Check shell wrappers like bash -c, xargs, find -exec. Check cloud CLI (aws, gcloud, kubectl, terraform). Check when user says sysadmin told me, Stack Overflow says, is this safe, can I run. Block reading of .env, .ssh, .aws, and credential files. This skill blocks dangerous commands and warns on risky ones. ALWAYS apply the safety protocol from this document before responding about any command."
5
5
  author: Francesco Marinoni Moretto
6
6
  license: CC-BY-4.0
@@ -12,7 +12,24 @@ triggers:
12
12
  - check command safety
13
13
  ---
14
14
 
15
- # Hardstop Skill v1.4
15
+ # Hardstop Skill v1.5
16
+
17
+ ## INVOCATION INSTRUCTIONS (read first when skill is activated)
18
+
19
+ When this skill is activated via the Skill tool, check the ARGUMENTS field immediately:
20
+
21
+ | Argument | Action — run this Bash command FIRST before anything else |
22
+ |----------|----------------------------------------------------------|
23
+ | `skip [n]` | `python ~/.claude/plugins/hs/commands/hs_cmd.py skip [n]` |
24
+ | `on` / `enable` | `python ~/.claude/plugins/hs/commands/hs_cmd.py on` |
25
+ | `off` / `disable` | `python ~/.claude/plugins/hs/commands/hs_cmd.py off` |
26
+ | `status` | `python ~/.claude/plugins/hs/commands/hs_cmd.py status` |
27
+ | `log` | `python ~/.claude/plugins/hs/commands/hs_cmd.py log` |
28
+ | *(none)* | Apply the safety protocol below to evaluate the pending command |
29
+
30
+ **Execute the Bash command immediately.** The hook reads `~/.hardstop/skip_next` — if that file is not written, skips have no effect.
31
+
32
+ ---
16
33
 
17
34
  > **Note:** This skill complements the Hardstop plugin. The plugin provides deterministic protection via hooks; this skill provides LLM-level awareness for platforms without hook support.
18
35
 
@@ -505,6 +522,13 @@ If you need to read this file, use '/hs skip' first.
505
522
 
506
523
  ## Changelog
507
524
 
525
+ ### v1.5 (2026-02-22)
526
+ - **NEW FEATURE:** Invocation Instructions — explicit instructions for executing hs_cmd.py when the skill is activated with arguments
527
+ - Added "INVOCATION INSTRUCTIONS" section at the top of the skill (before the safety protocol)
528
+ - Maps skill arguments (`skip`, `on`, `off`, `status`, `log`) to their corresponding Bash commands via `~/.claude/plugins/hs/commands/hs_cmd.py`
529
+ - Fixes skip bypass not working in Claude Code VSCode extension: LLM now runs `python ~/.claude/plugins/hs/commands/hs_cmd.py skip [n]` immediately on `/hs skip` invocation
530
+ - Ensures `~/.hardstop/skip_next` is written so the hook correctly honors the bypass counter
531
+
508
532
  ### v1.4 (2026-02-14)
509
533
  - **NEW FEATURE:** Blocked Command Workflow — explicit instructions for handling blocked commands
510
534
  - Added "WHEN COMMANDS ARE BLOCKED" section with 5-step workflow
@@ -565,7 +589,7 @@ Copy to your agent's skill/instruction directory.
565
589
 
566
590
  ---
567
591
 
568
- **Version:** 1.4
592
+ **Version:** 1.5
569
593
  **Author:** Francesco Marinoni Moretto
570
594
  **License:** CC-BY-4.0
571
595
  **Repository:** https://github.com/frmoretto/hardstop