2 Commits

Author SHA1 Message Date
hermes-pi 51ae214998 phronesis v0.4.0: bump project version for the linux-server-hardening skill
- per the owner's convention (2026-11-19): every added or modified
  bundled skill bumps the project version, so the security-skill
  port takes pyproject from 0.3.0 to 0.4.0 on the line the
  publish.py work already opened
- autotest fully green at 0.4.0

Generated by Mistral Vibe.
Co-Authored-By: Mistral Vibe <vibe@mistral.ai>
2026-11-19 11:53:10 +01:00
hermes-pi 0226e33ac7 phronesis v0.4.0-dev: add skill linux-server-hardening (security)
- port of the wiki-local linux-server-hardening checklist (v1.2.0) into
  the bundle: audit listeners and accounts, SSH key-only, close unused
  accounts, scoped firewall, fail2ban, unattended-upgrades, surface
  reduction, verify-and-record; new `security` category, announced live
  by the listing per CONTRIBUTING
- SKILL.md body ported verbatim from the source copy; front matter
  adapted to the four-key bundle contract (SPEC 4.2-4.3)
- MEMORIES.md starting template shipped beside it (SPEC 4.7); the
  installed memory is add-only so live instances keep their entries
  (SPEC 6.8)
- tests: bundled load, category listing, and category filter for the
  new skill; autotest fully green (84 tests, 93 percent coverage);
  e2e-verified via the installed pipx CLI

Generated by Mistral Vibe.
Co-Authored-By: Mistral Vibe <vibe@mistral.ai>
2026-11-19 11:50:41 +01:00
5 changed files with 404 additions and 1 deletions
+1
View File
@@ -64,6 +64,7 @@ phronesis/
| Skill | Version | Category | Purpose |
| --- | --- | --- | --- |
| [fiona](https://gitea.nolivaldeziii.com/nvaldez/phronesis/src/branch/master/skill/fiona/SKILL.md) | 2.2.0 | productivity | Operate an llm-wiki with the fiona-assistant CLI |
| [linux-server-hardening](https://gitea.nolivaldeziii.com/nvaldez/phronesis/src/branch/master/skill/linux-server-hardening/SKILL.md) | 1.2.0 | security | Numbered checklist to audit, harden, and verify an exposed Debian-family Linux server (SSH, accounts, firewall, fail2ban, updates) |
| [memory](https://gitea.nolivaldeziii.com/nvaldez/phronesis/src/branch/master/skill/memory/SKILL.md) | 1.0.1 | productivity | Give a skill its own MEMORIES.md beside its SKILL.md |
| [python-tdd](https://gitea.nolivaldeziii.com/nvaldez/phronesis/src/branch/master/skill/python-tdd/SKILL.md) | 1.2.0 | development | Minimalist TDD architect for Python work |
| [sectionize-markdown](https://gitea.nolivaldeziii.com/nvaldez/phronesis/src/branch/master/skill/sectionize-markdown/SKILL.md) | 1.2.0 | documentation | Number headings and list items so each rule can be cited |
+1 -1
View File
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
[project]
name = "phronesis"
version = "0.3.0"
version = "0.4.0"
description = "A chest of skills: list them, install them into an agent harness (vibe, ...)"
readme = "README.md"
requires-python = ">=3.11"
+173
View File
@@ -0,0 +1,173 @@
---
name: skill-memories
version: 1.0.1
title: linux-server-hardening Memories
created: 2026-11-19
updated: 2026-11-19
type: summary
---
# linux-server-hardening Memories
<!-- toc-start: headings 2..6 -->
- [1. How to use this file](#1-how-to-use-this-file)
- [1.1 When to read](#11-when-to-read)
- [1.2 When to write](#12-when-to-write)
- [1.3 How to write](#13-how-to-write)
- [1.4 How the memory evolves](#14-how-the-memory-evolves)
- [2. Output and wording](#2-output-and-wording)
- [2.1 Style and format](#21-style-and-format)
- [2.2 Placement and naming](#22-placement-and-naming)
- [3. Inputs and sources](#3-inputs-and-sources)
- [3.1 Where inputs come from](#31-where-inputs-come-from)
- [3.2 Quirks and parsing](#32-quirks-and-parsing)
- [4. Steps and tools](#4-steps-and-tools)
- [4.1 Commands and flags](#41-commands-and-flags)
- [4.2 Machine and shell](#42-machine-and-shell)
- [5. Recurring requests](#5-recurring-requests)
- [5.1 Patterns](#51-patterns)
<!-- toc-end -->
What users want from the linux-server-hardening skill, kept across sessions:
standing preferences, corrections, repeated request patterns, and
facts a run verified while using this skill. Preferences for the
whole wiki go to the wiki's `MEMORY.md`. Knowledge from sources goes
to pages under `wiki/`.
## 1. How to use this file
### 1.1 When to read
- **1.1.1** Read this file once [the skill](./SKILL.md) is loaded,
before its first step of the session. When it is already in
context, do not read it again.
- **1.1.2** Apply every entry that matches the request.
- **1.1.2.1** An entry overrides a default of the skill and an
entry of the wiki's `MEMORY.md`, for work done with this skill
only.
- **1.1.2.2** An entry never overrides the boundaries in the wiki's
`AGENTS.md`; when the two conflict, follow the boundary and ask
the user.
- **1.1.2.3** An entry never changes a tool contract: a command, a
flag, or a required frontmatter field.
- **1.1.3** Every heading and list item is numbered as
[sectionize-markdown](../../skills/sectionize-markdown/SKILL.md)
describes. Cite an item as `linux-server-hardening MEMORIES.md item 2.1.3`; a
new item takes the next free number in its subsection.
### 1.2 When to write
- **1.2.1** Write an entry when one of these happens while the skill
is in use:
- **1.2.1.1** The user states a requirement or preference for this
skill that outlasts the current request ("always", "never", "from
now on", "when you do X").
- **1.2.1.2** The user corrects the skill's output or a step it
took.
- **1.2.1.3** The same kind of request comes for the second time;
record the pattern and what the user wanted.
- **1.2.1.4** A run verifies a reusable fact about the skill's
inputs, hosts, or tools that the skill does not state.
- **1.2.1.5** The user explicitly asks you to remember something
for this skill.
- **1.2.2** Do not write:
- **1.2.2.1** One-off values that only apply to the current
request.
- **1.2.2.2** Passwords, tokens, cookies, personal data, or
anything the user asks to keep out.
- **1.2.2.3** Preferences that hold for any wiki work; those go to
the wiki's `MEMORY.md`.
- **1.2.2.4** Facts the skill or a wiki page already states; cite
them instead.
- **1.2.2.5** Guesses; only what the user said or a run showed.
### 1.3 How to write
- **1.3.1** Put the entry under the `###` subsection that fits.
- **1.3.1.1** When none fits, add a `###` subsection with the next
free number under the matching `##`, or a new `##` category with
the next free number.
- **1.3.1.2** Never put an entry directly under a `##`; its id
would collide with a later subsection.
- **1.3.2** Use the knowledge-entry shape of sectionize-markdown
section 4: a header line with the source tag and a short title,
one child per fact, and a last child with the ref and the date.
- **1.3.3** Source tags: `user` (the user said it), `run` (a command
or tool run showed it), `source` (a file under `raw/` states it),
joined with `+` when several agree.
- **1.3.4** The ref says where to check the fact: the user's request
and its date, the command and its output line, or a page or raw
path.
- **1.3.5** Write the entry in the same turn, then tell the user
which item was added or changed, for example "Saved to linux-server-hardening
MEMORIES.md item 2.1.1".
- **1.3.6** Bump `version` (patch for a changed entry, minor for a
new entry or subsection) and set `updated` in the frontmatter on
every change.
- **1.3.7** Append `- YYYY-MM-DD - linux-server-hardening memory: <item id> <title>`
to `wiki/log.md`, and bump that file too.
- **1.3.8** The change lands in the save of the current ingest cycle
or housekeeping pass.
- **1.3.8.1** With no cycle running, save it on its own with
`fiona-assistant git save --description "linux-server-hardening memory: <item id> <title>"`.
An entry looks like this:
```text
- **2.1.1** [user] "Reports open with the verdict"
- **2.1.1.1** The report starts with pass or fail, then the evidence.
- **2.1.1.2** Applies to every run of this skill.
- **2.1.1.3** Ref: user request YYYY-MM-DD; added YYYY-MM-DD.
```
### 1.4 How the memory evolves
- **1.4.1** Before adding an entry, look for a near-duplicate in the
same category; extend it instead and keep its id and its earliest
date.
- **1.4.2** A finding that contradicts an entry becomes its next
child, titled `OVERRIDE: ...`, with the new facts and its own ref.
- **1.4.3** An entry that no longer applies is marked
**Superseded YYYY-MM-DD** with the reason; it is not deleted.
- **1.4.4** An entry that held for several uses graduates; mark it
**Promoted YYYY-MM-DD** with the target.
- **1.4.4.1** A rule for every run of the skill moves into
[the skill](./SKILL.md).
- **1.4.4.2** A preference that turns out to hold for all wiki
work moves to the wiki's `MEMORY.md`.
- **1.4.5** Never renumber ids; pages, log lines, and chats may cite
them.
- **1.4.6** Git is the memory's history: `git log -p MEMORIES.md` in
the skill's folder shows when each entry changed and why.
## 2. Output and wording
What the skill produces and how it reads.
### 2.1 Style and format
### 2.2 Placement and naming
## 3. Inputs and sources
What the skill reads and where it comes from.
### 3.1 Where inputs come from
### 3.2 Quirks and parsing
## 4. Steps and tools
How the skill's commands and the machine behave.
### 4.1 Commands and flags
### 4.2 Machine and shell
## 5. Recurring requests
Requests that come back, and what the user wanted each time.
### 5.1 Patterns
+215
View File
@@ -0,0 +1,215 @@
---
name: linux-server-hardening
version: 1.2.0
description: Load this skill when a Linux server needs hardening before or after exposure to a network (Debian or Raspberry Pi homelab) - a numbered, human-followable checklist that audits listeners and accounts, enforces SSH key-only, closes unused accounts, scopes the firewall, sets up fail2ban and automatic security updates, and verifies every change with a real probe.
category: security
---
# Linux Server Hardening
<!-- toc-start: headings 2..6 -->
- [1. How to use this checklist](#1-how-to-use-this-checklist)
- [2. Audit the exposure](#2-audit-the-exposure)
- [3. Harden SSH](#3-harden-ssh)
- [4. Close unused accounts](#4-close-unused-accounts)
- [5. Firewall the ingress](#5-firewall-the-ingress)
- [6. Brute force protection](#6-brute-force-protection)
- [7. Automatic security updates](#7-automatic-security-updates)
- [8. Reduce the listening surface](#8-reduce-the-listening-surface)
- [9. Verify and record](#9-verify-and-record)
- [10. Memory](#10-memory)
<!-- toc-end -->
A hardening checklist for a Debian-family Linux server, written
so a human can follow it top to bottom. Every heading and list
item has a stable id; cite one as `item 3.4`. Items are facts to
hold and gotchas; boxes are steps to do and verify. The ground
rule throughout: a probe decides, memory and recollection do not.
## 1. How to use this checklist
- **1.1** Work in order 2 to 9: audit first (read-only), then
change, then verify. Never batch two changes between probes.
- **1.2** The skill assumes a Debian-family host with sudo; adapt
names and paths to the distribution in front of you.
- **1.3** Before editing any live config, copy it aside:
`sudo cp -a <file> /tmp/<file>.bak-$(date +%Y%m%d)`.
- **1.4** Edit live config in place; never reconstruct it from a
template or memory. Diff against the backup after the edit.
- **1.5** Every change gets a recorded one-line undo (the backup
path, the prior shell, the old rule set).
- **1.6** `sudo` is assumed on every command below; do not run a
shell as root.
## 2. Audit the exposure
Read-only; change nothing here. Build the picture before touching
anything.
- [ ] **2.1** Inventory the TCP and UDP listeners with their
owning processes:
`sudo ss -tlnp` and `sudo ss -ulnp`.
- [ ] **2.2** For every listener, name the service, why it runs,
and who may reach it (public, LAN, loopback). No unexplained
listener stays unexplained.
- [ ] **2.3** Inventory accounts that can log in:
`awk -F: '$7 ~ /(sh|bash)$/ {print $1":"$3}' /etc/passwd`.
- [ ] **2.4** Inventory accounts with a usable password:
`sudo awk -F: '$2 ~ /^\$/ {print $1}' /etc/shadow`.
- [ ] **2.5** Inventory sudo members: `getent group sudo`, and
check `authorized_keys` of every human account for content.
- [ ] **2.6** Record the firewall state: `sudo ufw status
verbose` (or the nftables equivalent).
- [ ] **2.7** Record pending updates: `apt list --upgradable
2>/dev/null | tail -n +2 | wc -l` and flag the security
ones.
- **2.8** A listener is not an open port: cross the listener list
with the firewall rules. "Open" means reachable from a given
source, and only the firewall says which sources those are.
## 3. Harden SSH
- **3.1** Keep everything in a drop-in, not the main file:
`/etc/ssh/sshd_config.d/10-no-password.conf` (included at the
top of `sshd_config` on Debian).
- [ ] **3.2** Write the drop-in with:
`PasswordAuthentication no`,
`KbdInteractiveAuthentication no`,
`PermitRootLogin no`, `MaxAuthTries 3`,
`X11Forwarding no`, `LoginGraceTime 30`.
- [ ] **3.3** Validate before activating:
`sudo sshd -t` must print nothing.
- [ ] **3.4** Activate gracefully: `sudo systemctl reload ssh`.
A reload drops no existing session; a restart can.
- [ ] **3.5** Verify the effective config, not the file:
`sudo sshd -T | grep -Ei '^(passwordauthentication|
kbdinteractiveauthentication|permitrootlogin)'`.
- [ ] **3.6** Verify behavior with a live probe:
`ssh -o PreferredAuthentications=none -o ConnectTimeout=5
localhost` must answer `Permission denied (publickey)`.
- **3.7** GOTCHA: if VNC or other services tunnel over SSH
(`ssh -L`), `AllowTcpForwarding` must stay `yes`. Turning it
off silently kills the tunnels and looks like "server not
responding".
- **3.8** GOTCHA: `PermitRootLogin without-password` (the
default) still allows root login with a key. An empty
`/root/.ssh/authorized_keys` saves you today; say `no` anyway.
- **3.9** Check the protocol only on museum-grade hosts: SSH
protocol 2 is enforced by the software itself on OpenSSH
7.6 and later - the `Protocol` directive was removed from
the config language and modern sshd cannot speak protocol 1.
Verify with `ssh -v` (`Remote protocol version 2.0`). A
"force protocol 2" item in a modern guide is a staleness
marker for the guide, not a gap in the host.
## 4. Close unused accounts
- [ ] **4.1** For each candidate account, verify first that
nothing uses it: `ps -eo user,args | grep <name>`, `sudo
crontab -l -u <name>`, enabled units, recent logins
(`sudo last | grep <name>`).
- [ ] **4.2** Lock the password: `sudo passwd -l <name>`.
- [ ] **4.3** Set the shell to `/usr/sbin/nologin`:
`sudo usermod -s /usr/sbin/nologin <name>`.
- [ ] **4.4** Verify: `sudo passwd -S <name>` shows `L`, and
`getent passwd <name>` shows the nologin shell.
- **4.5** GOTCHA: a password lock does NOT block SSH key login.
An account with content in `authorized_keys` stays reachable
until its shell is nologin (or the keys are removed).
- **4.6** Record the undo: `sudo passwd -u <name>` plus
`sudo usermod -s /bin/bash <name>` (or the prior shell).
## 5. Firewall the ingress
- [ ] **5.1** Set the policy: `sudo ufw default deny incoming`
and `sudo ufw default allow outgoing`.
- [ ] **5.2** Add allow rules scoped to a source, never to
`Anywhere` without a reason; write a comment per rule:
`sudo ufw allow from 192.168.178.0/24 to any port 22 proto
tcp comment 'SSH from LAN'`.
- [ ] **5.3** If access arrives through a proxy or jump host,
allow exactly that one IP for that port.
- [ ] **5.4** Enable and verify: `sudo ufw --force enable;
sudo ufw status verbose`. The current SSH session must
survive the enable; if it does not, you locked yourself
out.
- **5.5** Cover both stacks: repeat rules for the IPv6 prefixes
in use, and note that an ISP global prefix can rotate.
## 6. Brute force protection
- [ ] **6.1** Install and enable fail2ban with the sshd jail
(Debian default ships it); verify it counts probes:
`sudo fail2ban-client status sshd`.
- **6.2** GOTCHA: if public access arrives through a stream or
reverse proxy, every internet probe reaches sshd with the
proxy's IP. A ban can then only hit the proxy - cutting the
owner's own public path (a self-DoS an attacker can trigger)
while real attacker IPs stay invisible.
- [ ] **6.3** In that topology, add the proxy to ignoreip
(`/etc/fail2ban/jail.local`, `[DEFAULT]` section:
`ignoreip = 127.0.0.1/8 ::1 <proxy-ip>`), reload, and
verify with `sudo fail2ban-client get sshd ignoreip`.
- **6.4** Per-client banning in that topology belongs on the
proxy host, against its own (real-client-IP) logs. Do not
pretend it works on the backend box.
## 7. Automatic security updates
- [ ] **7.1** Install: `sudo apt-get install -y
unattended-upgrades`.
- [ ] **7.2** Enable: write `/etc/apt/apt.conf.d/20auto-upgrades`
with `APT::Periodic::Update-Package-Lists "1";` and
`APT::Periodic::Unattended-Upgrade "1";`.
- [ ] **7.3** Verify the timers are enabled:
`systemctl is-enabled apt-daily.timer
apt-daily-upgrade.timer`.
- [ ] **7.4** Verify the origins with `sudo unattended-upgrades
--dry-run --debug`: the distribution's security origin must
be in the allowed list.
- **7.5** GOTCHA: third-party repositories (vendor repos such as
the Raspberry Pi Foundation origin) are not in the default
allow list and stay manual; re-check `apt list --upgradable`
occasionally.
## 8. Reduce the listening surface
- [ ] **8.1** For every listener that survived the audit without
a reason, disable both the service and its socket:
`sudo systemctl disable --now <name>.service <name>.socket`
(example: rpcbind, when no NFS or rpc is in use).
- [ ] **8.2** Verify the port is gone from `sudo ss -tlnp` and
`sudo ss -ulnp`; a socket unit can re-activate a service.
- **8.3** Prefer not running a service over filtering it: a
disabled service is closed to every source, forever.
## 9. Verify and record
- [ ] **9.1** Re-run the probes of section 2; compare listener
by listener and rule by rule against the pre-hardening
picture.
- [ ] **9.2** Test the public path from a real outside leg (a
phone on mobile data), not from inside the LAN, and not
through a VPN.
- **9.3** GOTCHA: consumer VPN exits (verified with Mullvad)
silently drop non-web ports; a failed TCP test through such a
tunnel says nothing about the port. If a question remains, use
`https://check-host.net/check-tcp?host=<ip>:<port>` or a phone
on mobile data.
- [ ] **9.4** Record what changed and how to undo it on the
host's wiki entity page, then save the change set as one
commit. A hardening step that is not written down did not
happen.
## 10. Memory
- **10.1** After loading this skill, read [MEMORIES](./MEMORIES.md)
once per session. When it is already in context, do not read it
again.
- **10.2** Update it whenever an important memory about this skill
comes up: a lasting preference, a correction, a repeated
request, or a fact a run verified. Also update it whenever the
user explicitly asks you to remember something for this skill.
Section 1 of that file says how.
+14
View File
@@ -42,10 +42,24 @@ def test_skills_loads_the_bundled_python_tdd():
assert (skill.source_dir / "SKILL.md").is_file()
def test_skills_loads_the_bundled_linux_server_hardening():
assert "linux-server-hardening" in registry.SKILLS
skill = registry.SKILLS["linux-server-hardening"]
assert skill.version == "1.2.0"
assert skill.category == "security"
assert skill.description
assert (skill.source_dir / "SKILL.md").is_file()
assert (skill.source_dir / "MEMORIES.md").is_file()
def test_categories_list_the_packed_skills():
assert "python-tdd" in registry.CATEGORIES[registry.SKILLS["python-tdd"].category]
assert "linux-server-hardening" in registry.CATEGORIES["security"]
def test_skills_in_category_filters_by_label():
assert registry.skills_in_category("development") == {"python-tdd": registry.SKILLS["python-tdd"]}
assert registry.skills_in_category("security") == {
"linux-server-hardening": registry.SKILLS["linux-server-hardening"],
}
assert registry.skills_in_category("nonesuch") == {}