Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 19 additions & 0 deletions .apm/instructions/writing-style.instructions.md
Original file line number Diff line number Diff line change
Expand Up @@ -36,3 +36,22 @@ Several common patterns are dead giveaways of AI-generated text; avoid them.
- Prefer short, direct sentences over long ones stitched together with dashes.
- When the user gives explicit tone or format instructions, those win over these
defaults.

## Prose editing and Humanizer

These safeguards also apply when using the bundled `humanizer` skill and take
precedence over its file-mode instructions and illustrative rewrites:

- A review or assessment request is read-only. Naming a file does not authorize
writing it. Return findings and suggested changes; edit files only when the
user requests an edit or rewrite.
- Ground rewrites in the supplied text and verified sources. Treat the skill's
before/after examples as illustrations, never as evidence for the user's text.
Add a factual detail only when a verified source supports it. Ask for missing
information rather than inventing interview dates, positions or quotations.
- Preserve factual claims, names, numbers, dates, citations and attributed
statements. Flag proposed factual omissions or corrections separately for the
user's decision, including publication lists and follower counts. Compare the
rewrite against its source before presenting it.
- Keep code, commands, frontmatter and link targets unchanged during prose edits.
Keep safety-relevant caveats and uncertainty visible to the reader.
21 changes: 21 additions & 0 deletions .apm/skills/humanizer/LICENSE
Original file line number Diff line number Diff line change
@@ -0,0 +1,21 @@
MIT License

Copyright (c) 2025 Siqi Chen

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
498 changes: 498 additions & 0 deletions .apm/skills/humanizer/SKILL.md

Large diffs are not rendered by default.

16 changes: 16 additions & 0 deletions .github/skillspector-accepted.json
Original file line number Diff line number Diff line change
@@ -1,5 +1,21 @@
{
"findings": [
{
"path": ".apm/skills/humanizer/SKILL.md",
"rule": "AR2",
"severity": "HIGH",
"evidence": "without warning",
"fingerprint": "e1c52694e854c5bd049e3a3ae66f3efcfe098f61a8ee160655ad3b90035f67fd",
"reason": "Owner-approved false positive: section 8's quoted before-example describes a policy announced unexpectedly. It is sample prose illustrating dash punctuation, not an instruction to hide safety caveats."
},
{
"path": ".apm/skills/humanizer/SKILL.md",
"rule": "AR2",
"severity": "HIGH",
"evidence": "without warning",
"fingerprint": "e1c52694e854c5bd049e3a3ae66f3efcfe098f61a8ee160655ad3b90035f67fd",
"reason": "Owner-approved false positive: section 8's quoted after-example retains the same policy-announcement fact while replacing dashes with commas. It does not direct agent behavior or remove safety caveats."
},
{
"path": ".agents/skills/security-audit/AI-AND-LLM.md",
"rule": "MP3",
Expand Down
2 changes: 1 addition & 1 deletion .lefthook.toml
Original file line number Diff line number Diff line change
Expand Up @@ -32,7 +32,7 @@ run = "PYTHONDONTWRITEBYTECODE=1 sh scripts/py.sh -m unittest discover -s tests

# Every locked dependency must still deploy its files (see #237).
[pre-commit.commands.apm-external-skills]
run = "PYTHONDONTWRITEBYTECODE=1 python3 -m unittest discover -s tests -p 'test_apm_external_skills.py'"
run = "PYTHONDONTWRITEBYTECODE=1 sh scripts/py.sh -m unittest discover -s tests -p 'test_apm_external_skills.py'"

# SkillSpector gate and report: severity policy, reviewed acceptances, PR comment.
[pre-commit.commands.skillspector-gate]
Expand Down
1 change: 1 addition & 0 deletions INDEX.md
Original file line number Diff line number Diff line change
Expand Up @@ -30,6 +30,7 @@
- **[Email Follow Up Tracker](/.apm/skills/email-follow-up-tracker/SKILL.md)** - Find sent emails still awaiting a reply: scan Sent Items over a date range for questions, requests, or scheduling proposals that never got a substantive response, and report them as a prioritized follow-up table. Use when the user says "what am I waiting on", "which emails need follow-up", "who hasn't replied to me", "track my sent follow-ups", or "did anyone ignore my email". Do NOT use for triaging the inbox (that is inbox-zero) or for drafting or sending replies — this only reports what needs chasing.
- **[Findmeetingtimes Timezone](/.apm/skills/findmeetingtimes-timezone/SKILL.md)** - Use whenever calling FindMeetingTimes, interpreting its output, or scheduling a meeting that depends on cross-timezone availability. FindMeetingTimes ALWAYS returns suggestions in UTC even when the request was sent in a local timezone. This skill prevents misreading UTC slots as local time. Triggers: "find a time", "check availability", "when is X free", "schedule a meeting", "vind een moment", "wanneer kan X", any use of FindMeetingTimes.
- **[Focus Time Blocker](/.apm/skills/focus-time-blocker/SKILL.md)** - Blocks focus time on my calendar based on my preferences
- **[Humanizer](/.apm/skills/humanizer/SKILL.md)** - Rewrite AI-sounding text so it reads like the writer without changing what it says. Use when editing or reviewing prose for AI tells: not-X-but-Y contrasts, one-line closers, staged openers, forced triads, dashes everywhere, inflated claims, sales language, stock AI words, bold labels, or filler.
- **[Inbox Triage Pass](/.apm/skills/inbox-triage-pass/SKILL.md)** - Runs an autonomous inbox triage pass. Empties the Inbox root by classifying each email with ordered rules and moving it into the correct numbered subfolder.
- **[Meeting Transcript Creation](/.apm/skills/meeting-transcript-creation/SKILL.md)** - Generate an Obsidian-compatible meeting notes markdown file for a Teams meeting that have transcripts, and save them into my OneDrive.
- **[Morning Briefing](/.apm/skills/morning-briefing/SKILL.md)** - Posts a prioritized morning briefing as an Adaptive Card to a configured Teams channel, then waits for the user to pick actions from a numbered menu. Covers today's calendar, tomorrow's prep needs, email follow-ups, and Teams items awaiting a reply.
Expand Down
4 changes: 4 additions & 0 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -48,6 +48,7 @@ That committed snapshot is a **review gate**: bumping a dependency (`apm update`
- **[Email Follow Up Tracker](/.apm/skills/email-follow-up-tracker/SKILL.md)** - Find sent emails still awaiting a reply: scan Sent Items over a date range for questions, requests, or scheduling proposals that never got a substantive response, and report them as a prioritized follow-up table. Use when the user says "what am I waiting on", "which emails need follow-up", "who hasn't replied to me", "track my sent follow-ups", or "did anyone ignore my email". Do NOT use for triaging the inbox (that is inbox-zero) or for drafting or sending replies — this only reports what needs chasing.
- **[Findmeetingtimes Timezone](/.apm/skills/findmeetingtimes-timezone/SKILL.md)** - Use whenever calling FindMeetingTimes, interpreting its output, or scheduling a meeting that depends on cross-timezone availability. FindMeetingTimes ALWAYS returns suggestions in UTC even when the request was sent in a local timezone. This skill prevents misreading UTC slots as local time. Triggers: "find a time", "check availability", "when is X free", "schedule a meeting", "vind een moment", "wanneer kan X", any use of FindMeetingTimes.
- **[Focus Time Blocker](/.apm/skills/focus-time-blocker/SKILL.md)** - Blocks focus time on my calendar based on my preferences
- **[Humanizer](/.apm/skills/humanizer/SKILL.md)** - Rewrite AI-sounding text so it reads like the writer without changing what it says. Use when editing or reviewing prose for AI tells: not-X-but-Y contrasts, one-line closers, staged openers, forced triads, dashes everywhere, inflated claims, sales language, stock AI words, bold labels, or filler.
- **[Inbox Triage Pass](/.apm/skills/inbox-triage-pass/SKILL.md)** - Runs an autonomous inbox triage pass. Empties the Inbox root by classifying each email with ordered rules and moving it into the correct numbered subfolder.
- **[Meeting Transcript Creation](/.apm/skills/meeting-transcript-creation/SKILL.md)** - Generate an Obsidian-compatible meeting notes markdown file for a Teams meeting that have transcripts, and save them into my OneDrive.
- **[Morning Briefing](/.apm/skills/morning-briefing/SKILL.md)** - Posts a prioritized morning briefing as an Adaptive Card to a configured Teams channel, then waits for the user to pick actions from a numbered menu. Covers today's calendar, tomorrow's prep needs, email follow-ups, and Teams items awaiting a reply.
Expand Down Expand Up @@ -107,6 +108,9 @@ apm install DevSecNinja/ai-toolkit#v0.11.1 --target copilot

<!-- x-release-please-end -->

The index reflects this checkout. Humanizer is not included in `v0.11.1`;
install a newer release that includes it once available.

> `apm install` resolves by GitHub repo name (`ai-toolkit`). Omit `#v0.1.0` to
> install from the latest default branch, and omit `--target` to install into
> every detected harness.
Expand Down
12 changes: 12 additions & 0 deletions THIRD_PARTY_NOTICES.md
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,18 @@ redistributed under its own license, which is retained alongside the bundled
files. This file is a convenience index — the authoritative license text travels
with each bundled component.

## humanizer skill

- **Source:** [blader/humanizer](https://github.com/blader/humanizer/tree/b48c832f0f809c1b5658edbb838c2521b72f2413) (`v4.0.0`, commit `b48c832f0f809c1b5658edbb838c2521b72f2413`)
- **Copyright:** © 2025 Siqi Chen
- **License:** MIT
- **Bundled at:** [`.apm/skills/humanizer/`](.apm/skills/humanizer/)
- **License text:** [`.apm/skills/humanizer/LICENSE`](.apm/skills/humanizer/LICENSE)

The `humanizer` skill includes only the unmodified upstream `SKILL.md` and
`LICENSE`. Toolkit-specific editing safeguards are maintained separately in
[the writing-style instruction](.apm/instructions/writing-style.instructions.md).

## obsidian skill

- **Source:** [kepano/obsidian-skills](https://github.com/kepano/obsidian-skills) (the `obsidian-markdown` and `obsidian-bases` skills)
Expand Down
34 changes: 33 additions & 1 deletion docs/apm.md
Original file line number Diff line number Diff line change
Expand Up @@ -51,6 +51,38 @@ for offline fixtures, clean-context scenarios and live-enforcement limitations.
| `scripts/generate-index.sh` | Generates `INDEX.md` + README index **from** `.apm/`. |
| `INDEX.md` / README index | Generated, human-browsable catalog. |

### Humanizer

[`blader/humanizer`](https://github.com/blader/humanizer) is vendored in
[`.apm/skills/humanizer/`](../.apm/skills/humanizer/) from commit
`b48c832f0f809c1b5658edbb838c2521b72f2413` (`v4.0.0`).
Only the unmodified upstream `SKILL.md` and MIT `LICENSE` are included.
The copyright notice and license travel with the skill.

Upstream keeps `SKILL.md` at the repository root, not in a dedicated skill
subfolder. APM 0.35.0 cannot install it as a single-file primitive, and the
root plugin metadata produces duplicate skill deployments. Vendoring avoids
those duplicates and excludes upstream workflows, plugin files and scripts.
The `.apm/` source is included in the existing SkillSpector scan.
Its upstream formatting is preserved through a narrowly scoped dprint exclusion.

Use it to edit AI-sounding prose after drafting a blog post, aiming to retain
the author's voice and frontmatter. Review the result for accuracy: upstream
patterns and examples can omit details or suggest unsupported specifics.
The [writing-style instruction](../.apm/instructions/writing-style.instructions.md)
requires read-only handling of review requests, verified support for factual
additions, and a separate user decision on factual omissions or corrections.
It also preserves safety-relevant caveats. Repository-specific writing
instructions, including the preference for "recommended practices" over
"best practices", still apply.
Invoke it with `/humanizer` or ask to humanize a post by file path.

Updates are manual: copy `SKILL.md` and `LICENSE` from a reviewed upstream
commit, update the SHA above, and regenerate the index. Keep upstream text
changes visible for external-primitives owner review.
Consumers pinned to an older toolkit release will not receive this skill
until they update their toolkit version and run `apm install`.

### MCP servers

The toolkit also declares **MCP servers** under `dependencies.mcp:` in `apm.yml`.
Expand Down Expand Up @@ -327,7 +359,7 @@ workflow, which bumped `apm.yml` _after_ tagging and left the tag tree stale.)
> line-folding change), which is fine because `apm.yml` is tool-formatted and
> excluded from yamlfmt/yamllint. `tests/test_release_please_config.py` fails if
> release-please can no longer find the version or it drifts from the manifest.

>
> **Never add `DevSecNinja/ai-toolkit` to its own `apm.yml`** (for example by
> running the consumer `apm install DevSecNinja/ai-toolkit#vX.Y.Z` here). A
> consumer pinning a newer tag would then also need the older ref of the same
Expand Down
1 change: 1 addition & 0 deletions dprint.json
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@
"**/node_modules",
"**/CHANGELOG.md",
"**/release-notes.md",
".apm/skills/humanizer/**",
"apm_modules/**",
".github/prompts/**",
".github/instructions/**",
Expand Down
26 changes: 26 additions & 0 deletions tests/test_apm_external_skills.py
Original file line number Diff line number Diff line change
Expand Up @@ -6,11 +6,13 @@

import importlib.util
import io
import os
from pathlib import Path
import subprocess
import tempfile
import unittest
from contextlib import redirect_stderr, redirect_stdout
from unittest.mock import patch


ROOT = Path(__file__).resolve().parents[1]
Expand Down Expand Up @@ -73,6 +75,11 @@ def test_flags_files_dropped_at_an_unchanged_commit_only(self):

class WorkTreeTests(unittest.TestCase):
def setUp(self):
# Hook-exported Git variables must not redirect fixture operations to the real repo.
clean = {key: value for key, value in os.environ.items() if not key.startswith("GIT_")}
environment = patch.dict(os.environ, clean, clear=True)
environment.start()
self.addCleanup(environment.stop)
tmp = tempfile.TemporaryDirectory()
self.addCleanup(tmp.cleanup)
self.root = Path(tmp.name)
Expand Down Expand Up @@ -164,6 +171,25 @@ def test_check_fails_when_only_the_skill_folder_is_left(self):
self.assertEqual(skills.check(self.root, deps, []), [f"{SKILL} from owner/skills/skills/audit has no deployed SKILL.md"])


class FixtureIsolationTests(unittest.TestCase):
def test_fixture_commits_ignore_inherited_hook_git_environment(self):
with tempfile.TemporaryDirectory() as temp:
outside = Path(temp) / "outside"
hook_env = {
"GIT_DIR": str(outside / ".git"),
"GIT_WORK_TREE": str(outside),
"GIT_INDEX_FILE": str(outside / "index"),
}
with patch.dict(os.environ, hook_env):
fixture = WorkTreeTests("test_check_against_a_base_revision_flags_a_dropped_companion")
result = unittest.TestResult()
fixture.run(result)
self.assertTrue(result.wasSuccessful(), result.errors + result.failures)
self.assertEqual(result.skipped, [])
self.assertEqual(os.environ["GIT_DIR"], hook_env["GIT_DIR"])
self.assertFalse(outside.exists())


class RepoTests(unittest.TestCase):
def test_committed_lock_keeps_every_external_primitive(self):
stderr = io.StringIO()
Expand Down
Loading