respond-compromised-account
Respond to a potentially compromised user account. Use when impossible travel, credential stuffing, successful phishing, or suspicious activity indicates account compromise. Investigates activity, contains the account, removes persistence, and restores access.
84
81%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Passed
No known issues
Quality
Discovery
100%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is an excellent skill description that clearly defines its purpose, lists concrete actions, and provides explicit trigger conditions via a well-constructed 'Use when' clause. The trigger terms are natural and specific to the account compromise domain, making it highly distinguishable from other skills. It follows the third-person voice convention and is concise without being vague.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: 'Investigates activity, contains the account, removes persistence, and restores access.' These are clear, actionable steps in an incident response workflow. | 3 / 3 |
Completeness | Clearly answers both 'what' (investigates activity, contains the account, removes persistence, restores access) and 'when' (explicit 'Use when' clause listing impossible travel, credential stuffing, successful phishing, or suspicious activity). | 3 / 3 |
Trigger Term Quality | Includes highly natural trigger terms that a user/analyst would actually use: 'impossible travel', 'credential stuffing', 'successful phishing', 'suspicious activity', 'account compromise', 'compromised user account'. These cover the main scenarios well. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive with a clear niche around compromised user account response. The specific trigger terms like 'impossible travel', 'credential stuffing', and 'phishing' make it unlikely to conflict with other security or general IT skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
62%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a solid incident response playbook with a well-structured PICERL workflow, clear decision matrices, and appropriate safety gates requiring analyst confirmation before destructive actions. Its main weaknesses are that several critical steps (identity provider actions, email platform actions) lack concrete tool calls, and the document is lengthy enough that it would benefit from splitting detailed sub-procedures into referenced files. Some content is mildly redundant between the phase descriptions and summary tables.
Suggestions
Provide concrete tool calls or explicit command examples for identity provider actions (disable account, reset password, revoke sessions) rather than noting they require IDP tools—or clearly document which specific MCP tool/server handles these.
Split detailed sub-procedures (endpoint triage, OAuth app investigation, email forwarding rule removal) into linked reference files to reduce the main skill's token footprint.
Remove the redundant 'Common Persistence Mechanisms' table at the end since the same information is already covered in Phase 4 Steps 4.2-4.3, or consolidate into one location.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is reasonably efficient but has some redundancy—e.g., persistence mechanisms are listed in both Phase 4 and the 'Common Persistence Mechanisms' table, and some items in the eradication/recovery phases repeat what Claude would naturally infer. The output tables at the top add bulk that could be trimmed. | 2 / 3 |
Actionability | Provides concrete tool calls (secops-mcp.search_security_events, secops-mcp.lookup_entity) and slash commands, which is good. However, many steps rely on tools marked as '*(Requires Identity Provider tools)*' or '*(Requires email/cloud platform tools)*' without specifying actual commands, making those steps vague. Containment and eradication execution steps are descriptive rather than executable. | 2 / 3 |
Workflow Clarity | The PICERL phased workflow is clearly sequenced with explicit steps, a containment decision matrix based on compromise likelihood, confirmation gates before destructive actions (/confirm-action), verification steps after containment (Step 3.3), and documentation checkpoints at each phase. The feedback loop of verify-then-proceed is well-established. | 3 / 3 |
Progressive Disclosure | The content is well-structured with clear headers and phases, but it's a long monolithic document (~200+ lines) with no references to external files for detailed sub-procedures (e.g., endpoint triage, OAuth app revocation). The reference tables at the end help but the overall length could benefit from splitting detailed procedures into linked files. | 2 / 3 |
Total | 9 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
frontmatter_unknown_keys | Unknown frontmatter key(s) found; consider removing or moving to metadata | Warning |
Total | 10 / 11 Passed | |
- Repository
- dandye/ai-runbooks
- Commit
086cbf6
- Reviewed
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.