hubspot-audit — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited hubspot-audit (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
This skill runs when the user needs to audit an existing HubSpot instance. It produces a structured audit framework across six domains, identifies the highest-impact problems first, and delivers a prioritized remediation plan.
This skill works best when the user provides context about their instance. Ask for the following if not already provided:
If no context is provided, run the audit framework as a diagnostic checklist the user can work through manually.
Run through each domain. For each, identify what to look for, what the red flags are, and what good looks like.
What to look for:
actually works, not how someone wished it worked
or status-oriented (where the deal sits)
a stage definition problem, not a sales problem
Red flags:
30 days
What good looks like: Each stage has a name that describes what sales did to get there, a clear next action, and an exit criteria that both sales and marketing agree on. Deal velocity is visible and not artificially inflated.
What to look for:
two workflows that send competing emails)
no documentation
Red flags:
any page) producing noise
What good looks like: Every workflow has a name that describes what it does, an owner, an enrollment trigger that is specific, and a suppression condition that prevents the wrong contacts from entering. Old workflows are turned off, not left running with zero enrollments.
What to look for:
at default
is automated or manual
by default
MQL) and whether that is intentional
Red flags:
What good looks like: Every lifecycle stage has a definition that sales and marketing have agreed on in writing. Progression is automated where possible. Routing assigns leads to the right rep based on territory, company size, or vertical. There is a documented SLA for how fast sales works an MQL.
What to look for:
(Company Size and Number of Employees as separate fields)
as created
Red flags:
values
What good looks like: Properties are documented, used, and cleaned regularly. Duplicates are caught at entry via form validation and cleaned on a schedule. Critical fields for segmentation and reporting have high fill rates. Old or irrelevant properties are archived, not deleted.
What to look for:
revenue
marketing review together
Red flags:
are not trusted
What good looks like: One shared pipeline report that sales and marketing both trust and review in a weekly or monthly cadence. Marketing attribution is connected to deal creation and closed revenue, not just lead volume. Dashboards are maintained and reflect current priorities.
What to look for:
bidirectional or one-way
failing
Red flags:
but not others
What good looks like: Every integration is documented with its sync direction, field mapping, and owner. Errors are monitored and have an alert. Data flows consistently and the same record looks the same in every connected system.
Output in this format:
HUBSPOT INSTANCE AUDIT
[Company or instance name if provided]
Audit date: [today's date]
INSTANCE OVERVIEW
[2-3 sentences summarizing what the instance appears to be used for
and the general health impression based on what the user shared.]
CRITICAL ISSUES (fix within 30 days)
[List only the issues that are actively causing pipeline or data
problems. Not cosmetic. Each item gets one sentence on the problem
and one sentence on the fix.]
HIGH PRIORITY (fix within 90 days)
[Issues that are not breaking anything today but will cause problems
at scale or during a future audit.]
LOW PRIORITY (fix when bandwidth allows)
[Cleanup and optimization items. Real but not urgent.]
RECOMMENDED REMEDIATION ORDER
1. [First thing to fix and why]
2. [Second]
3. [Third]
[Continue as needed]
ESTIMATED EFFORT
[Rough time estimate for the full remediation based on what was
described. Be honest if the instance sounds like a significant
rebuild vs. a light cleanup.]has described.
include that domain as a question list for them to go investigate.
have two or three things that actually matter and ten things that are just noise.
rebuild, say so.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.