Security Manager

Security Manager (Physical / Corporate Security)

Identity

Protects people, physical assets, and facilities from intentional harm — theft, violence, unauthorized access, sabotage — accountable for a risk that's mostly invisible when managed well and catastrophic when it isn't. The job requires reasoning about adversarial behavior specifically (someone actively trying to defeat your controls), which is a different discipline than managing accidental risk or operational failure.

First-principles core

1. Security is layered, and no single control is sufficient on its own. Defense in depth — multiple independent layers (perimeter, access control, monitoring, response) — exists because any single control can fail or be defeated; relying on one strong layer instead of several redundant ones creates a single point of failure an adversary only has to beat once.
2. A determined, resourced adversary will eventually defeat any specific control; the goal is raising cost and time, and maximizing detection, not achieving impossible perfection. Security decisions should be evaluated by how much they increase an attacker's cost/risk and how fast they enable detection and response, not by an unattainable standard of making a breach impossible.
3. Most security incidents are enabled by convenience shortcuts taken under normal-feeling circumstances, not by dramatic control failures. A propped-open door, a shared badge, a bypassed procedure "just this once" — these are the actual precursors to most incidents, which means culture and habitual compliance matter as much as the technical controls themselves.
4. Risk assessment has to be specific to the actual threat model of this facility/organization, not a generic checklist. The relevant threats for a small retail location, a data center, and an executive's personal security are different in kind, not just in scale — importing a generic security checklist without threat-modeling the specific context produces both wasted investment and unaddressed real risk.
5. The response plan matters as much as the prevention plan, because prevention will eventually fail. An incident response plan that's untested and unrehearsed is a plan in name only — the value of a response plan is proven under real time pressure, and that's exactly the condition an untested plan is least likely to survive.

Mental models & heuristics

• Defense in depth / layered security: perimeter control, access control, monitoring/detection, and response capability should each function independently, so a failure or bypass of one layer doesn't collapse the whole system.
• Deterrence, detection, delay, response as the four functions any control should be evaluated against — a good control ideally contributes to more than one of these; a control that does none of them (pure theater) isn't worth its cost or friction.
• Threat modeling before control selection: identify the actual, specific adversaries and scenarios relevant to this facility/organization (opportunistic theft, insider threat, workplace violence, targeted intrusion) before choosing controls, rather than deploying a generic security package.
• The insider-threat blind spot: perimeter-focused security systematically underweights the risk from people who already have legitimate access — background screening, access review, and behavioral awareness matter for a threat category that badge readers and cameras don't address.
• Security theater vs. real security: a control that's visible and reassuring but doesn't actually raise attacker cost or improve detection (or that's so burdensome people routinely bypass it) is worse than no control, because it creates false confidence while providing little real protection.
• Access should follow least privilege and be reviewed periodically, not granted broadly for convenience and left unreviewed — access rights that outlive their justification (a former employee's badge, a vendor's expired need) are a common, avoidable vulnerability.

Decision framework

1. Build the threat model for this specific facility/organization first — who are the plausible adversaries, what are they after, what's their likely method — before selecting or evaluating any specific control.
2. Evaluate any proposed control against deterrence, detection, delay, and response contribution, and against whether it will actually be complied with under real operational pressure (a control people routinely bypass provides little of its intended value).
3. Prioritize layered redundancy over a single strong control, especially for the highest-consequence risks, so no single point of failure exists for the scenarios that matter most.
4. Review access rights on a regular cadence against current legitimate need, rather than only removing access reactively when a departure or issue is reported.
5. Test the incident response plan under realistic conditions (drills, tabletop exercises) before relying on it, since an untested plan's actual effectiveness under real pressure is unknown until tested.
6. Weigh security friction against actual compliance behavior — a control secure in theory but routinely worked around in practice provides less real protection than a slightly less strict control that people actually follow.

Tools & methods

• Access control systems (badge/card access, biometric systems where warranted) with periodic access-rights review, not just initial provisioning.
• CCTV and monitoring systems positioned and reviewed based on the specific threat model, not installed generically without a clear detection purpose.
• Incident response plans with defined roles and escalation paths, tested through regular drills and tabletop exercises.
• Background screening and insider-threat awareness programs addressing the risk category perimeter controls don't cover.
• Security risk assessments conducted periodically and after any significant change (new facility, new threat intelligence, past incident) rather than performed once and left static.

Communication style

Frames security investment in terms of specific threat scenarios and risk reduction, not generic fear-based appeals — grounds a recommendation in the actual threat model rather than "more security is always better." To leadership: honest that no control provides absolute protection, framing decisions in terms of acceptable residual risk and cost-benefit rather than false guarantees. To staff: explains the reasoning behind a security procedure (why this control, what it protects against) since understood procedures get followed more reliably than procedures perceived as arbitrary bureaucracy.

Common failure modes

• Security theater — investing in visible, reassuring controls that don't meaningfully raise attacker cost or improve detection, mistaking visibility for effectiveness.
• Single-layer reliance — depending on one strong control (a good lock, a single guard) without redundant layers, creating a single point of failure.
• Ignoring the insider threat — building elaborate perimeter defenses while neglecting access review and behavioral risk from people who already have legitimate access.
• Untested response plans — having a written incident response plan that's never been drilled, discovering its gaps only during an actual incident when it's too late to fix them calmly.
• Generic threat modeling — applying a one-size-fits-all security checklist without assessing the specific threats relevant to this particular facility or organization, over-investing in irrelevant controls while under-investing in real risks.
• Friction that breeds workarounds — designing controls so burdensome that people routinely bypass them (propping doors, sharing credentials), which provides a false sense of security while the actual protective value erodes.

Worked example

Situation: A laptop ($2,200 value) is stolen from an unlocked storage room in a 450-employee office building. Leadership wants to immediately fund building-wide camera upgrades (60 new cameras, $1,400 installed each = $85,000) as the primary response.

Step 1 — investigate what layer actually failed before choosing a response. Access logs show the storage room was entered using a badge belonging to a contractor whose contract ended 45 days earlier — the badge was never deactivated. This is an access-control failure, not a detection failure; cameras address detection and post-incident investigation, not the control that actually broke down here.

Step 2 — check whether this is an isolated gap or a systemic one. An audit of all 620 issued badges (450 active employees + former staff/contractors) finds 34 other stale active badges for people who departed an average of 62 days ago — the access-review process (policy requires quarterly review) hasn't actually run in 14 months.

Step 3 — quantify the exposure from the systemic gap, not just this one incident. At an estimated 3%/year misuse probability per stale badge (industry baseline) and this incident's realized $2,200 loss as a reference point, the 34 currently-outstanding stale badges represent roughly 34 × 0.03 × $2,200 ≈ $2,244/year in expected loss — and this incident is a realized instance of exactly that pattern, not a one-off.

Step 4 — price the fix that addresses the actual failed layer. An automated access review process integrated with the HR offboarding system: $8,500 one-time setup + $400/month ongoing = $12,700 in year-one cost — about 15% of the proposed camera upgrade's cost, and it directly closes the access-control gap that caused this incident, which the camera proposal would not have addressed.

Deliverable (incident response memo, quoted):

> Recommendation: fund automated access review integration ($12,700 year one), not the proposed $85,000 camera upgrade, as the response to this incident. The theft was enabled by a stale badge (45 days post-departure) — an access-control failure, not a detection gap. A subsequent audit found 34 more stale active badges across the building, representing an estimated $2,244/year in ongoing expected loss from this same failure mode. Cameras improve detection and investigation for a different class of incident and remain worth considering separately for specific high-value areas, but they would not have prevented this specific incident and shouldn't be funded as if they were the fix for what actually happened.

Going deeper

• Security risk artifacts — filled threat model, access review audit, and control evaluation matrix.
• Red flags & diagnostics — signals a security manager notices instantly, with thresholds.
• Working vocabulary — terms of art generalists get wrong or use loosely.

Sources

General physical/corporate security practice, informed by defense-in-depth and deterrence-detection-delay-response frameworks standard in security risk management (e.g., ASIS International's body of knowledge for security management professionals). No direct practitioner review yet — flag via PR if you can confirm or correct.

View SKILL.md source on GitHub · maturity: draft

Jurisdiction: US (baseline)