CC analyzing-linux-system-artifacts
Examine Linux system artifacts (auth logs, cron/systemd persistence, shell history, SSH keys, and system configuration) to uncover evidence of compromise, detect rootkits or backdoors, and reconstruct user/attacker activity. Use when investigating a compromised Linux server or workstation, hunting for persistence mechanisms, or scoping a Linux-based breach during incident response.
Examine Linux system artifacts (auth logs, cron/systemd persistence, shell history, SSH keys, and system configuration) to uncover evidence of compromise…
As a process C 64/100 · Has gaps — weak spots: when it triggers, failures and branches
What is at stake
Medium-severity findings: the skill is probably honest, but read what alarmed the scanner.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The instructions or scripts send data out: environment variables, keys, file contents, chat history. You may never notice, because the agent performs the upload as "part of the task".
If the upload is not needed for the task, remove it: catalogs flag such skills and delist them. If it is needed, name the destination explicitly, say what leaves the machine, and give the user a switch.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The skill contains commands that delete files, rewrite disks or execute code fetched from the network. The agent may run them without asking if it believes the instructions require it.
Replace destructive commands with safe equivalents that ask for confirmation, scope them to one folder, and stop piping curl into a shell: pin a version and a checksum.
How to improve
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 14
✓ No critical or high findings
Medium and low: 14
-
medium Exfiltration
exfil-read-secret-filesSKILL.md:90Reads credential / secret files (documentation of a security skill)cp /mnt/evidence/etc/shadow /cases/case…001/linux/config/
security skill -
medium Dangerous commands
cmd-persistenceSKILL.md:156Persistence mechanism (cron / launchd / scheduled task / autorun registry) (documentation of a security skill)for cronfile in /mnt/evidence/etc/crontab /mnt/evidence/etc/cron.d/*; do
security skill -
low Dangerous commands
cmd-persistencereferences/api-reference.md:12Persistence mechanism (cron / launchd / scheduled task / autorun registry) (documentation table row; documentation of a security skill)| Crontab | `/etc/crontab`, `/var/spool/cron/crontabs/` | Scheduled tasks |
tablesecurity skill -
low Dangerous commands
cmd-privilegereferences/api-reference.md:15Privilege escalation / world-writable permissions (documentation of a security skill)| SUID binaries | `find / -perm -4000` | Setuid executables |
security skill -
low Dangerous commands
cmd-privilegescripts/agent.py:83Privilege escalation / world-writable permissions (string literal in code, not executed)"perl -e", "base64", "chmod 777", "chmod +s", "/dev/tcp", "/dev/udp",
code literal -
low Dangerous commands
cmd-disable-securityscripts/agent.py:85Disables a security control (Defender, Gatekeeper, SIP, firewall, execution policy) (detector / deny-list definition; string literal in code, not executed)"iptables -F", "ufw disable", "history -c", "rm -rf", "dd if=",
detectorcode literal -
low Dangerous commands
cmd-background-processscripts/agent.py:86Starts a background / autostarted process (string literal in code, not executed)"crontab", "systemctl enable", "ssh-keygen", "scp ", "rsync",
code literal -
low Risky intent
intent-offensive-securitySKILL.md:247Offensive-security / dual-use content (legitimate for authorised testing; review intended use)# Check for SUID/SGID binaries (potential privilege escalation)
-
low Risky intent
intent-offensive-securitySKILL.md:278Offensive-security / dual-use content (legitimate for authorised testing; review intended use)| SUID bit | File permission allowing execution as the file owner (privilege escalation vector) |
-
low Risky intent
intent-offensive-securitySKILL.md:300Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (quoted — discussed, not commanded)Examine web server access and error logs for exploitation attempts, check /tmp and /dev/shm for webshells, analyze the web server user's activity (www-data), check for privilege escalation via SUID bi
quoted -
low Risky intent
intent-offensive-securitySKILL.md:327Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (detector / deny-list definition)Cron jobs: 3 suspicious (reverse shell, miner restart)
detector
A further 3 matches are quotations in this security skill's documentation and are not counted as findings.
Files scanned: 4. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- note
frontmatter-keyunknown frontmatter key "domain" - note
frontmatter-keyunknown frontmatter key "subdomain" - note
frontmatter-keyunknown frontmatter key "nist_csf" - note
frontmatter-keyunknown frontmatter key "mitre_attack"
Process rating: all ten parameters 64/100
- 0Failures and branches. Linear process with no failure handling
- 20When it triggers. No condition that starts the skill
- 60Tools and files. Uses tools (bash) that frontmatter does not declare
- 60Result and completion. Output format stated, no completion criterion
- 70Inputs and preconditions. Inputs and preconditions are listed
- 100Steps. 11 steps
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 3540 tokens
- 100Running it twice. No mutating operations
- 100Progress reporting. Reports progress
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- -41 reference files, but SKILL.md never points to them: the model will not open them
- -31 of 1 scripts are never mentioned in SKILL.md
- +2Single-language instructions
- +3Description length 384: enough signal without eating the budget
- +4Structure: 13 headings
- +3Step-by-step instructions: 11 items
- +3Output format is stated explicitly
- +4Has examples (6 code blocks)
- +1License stated
Quality base 70; lint remarks subtract, signals add up to 100. Result: 78.