Anxiety for Highly Competent People — 100-day Protocol for Engineers Who Have Impostor Syndrome
You're good at your job but you're convinced everyone will realize you don't belong. Your anxiety isn't about being competent. It's about being seen as competent.
This is the core protocol. Over 100-Day we build the evidence store, ship the weekly retro, and land the patch below — until a rough code review stops escalating to a P1.
normative summary: instrument → store evidence → diff anxiety against reality, weekly.
- Request for Comments
- 2026
- Category
- Standards Track
- Vertical
- V5 · anxiety & calm
- Avatar
- A3 · engineers and PMs
- Program
- 100-Day
- Status
- PROPOSED STANDARD
Three reproducible defects in the threat model. If any reproduce on your system, this document is addressed to you.
1 · Problem Statement
Your anxiety is not a character flaw. It is a monitoring system with its severity thresholds set catastrophically wrong — paging P1 for events that merge clean.
- 1.1P1IMPOSTOR-001
Every green build still reads as fragile
You ship working code and immediately think 'this is fragile, someone will find a critical flaw, they'll see I'm a fraud.' You can't internalize success. Each win feels temporary.
- 1.2P1IMPOSTOR-002
One bad signal escalates straight to “career over”
You catastrophize: one bad code review = they're about to fire you. A critical bug = your career is over. You know rationally these are overreactions but your nervous system is screaming danger.
- 1.3P2IMPOSTOR-003
Perfectionism, running as a denial-of-service on you
You're trying to be perfect to prove you belong. You work 60-hour weeks. You're exhausted, stressed, anxious. And still convinced it's not enough. The treadmill never stops.
Three normative moves — the same review loop you already run on prod, pointed at the system that pages you about prod.
2 · Specification
The key words MUST, SHALL, and SHOULD are to be interpreted as in a real spec: non-negotiable, required, and strongly recommended.
- §2.1MUSTsplit_signal
Separate performance anxiety from actual performance
We track: What did you actually accomplish? vs. What did anxiety tell you about it? These are usually wildly different. Your anxiety says 'you'll get fired.' Reality says 'solid Q2, on track for promotion.'
- §2.2SHALLevidence_store
Build an evidence system
Every time you complete a project, we document: scope, deadline, quality, impact. This becomes your personal evidence library. When anxiety says 'you're incompetent,' you point to 50 successful projects. Math beats anxiety.
- §2.3SHOULDweekly_retro
Friday retro: anxiety vs. reality
What did anxiety predict? What actually happened? How accurate was anxiety? Usually: anxiety is wrong 80%+ of the time. Your nervous system learns: 'False alarm, you're actually fine.'
The core patch. Over 100-Day this diff is applied, reviewed, and merged for real: the threat model stops compiling “they are about to fire me” from a single bad signal.
3 · The Patch
the prediction
what anxiety swore would happen
the evidence
what the record actually shows
the anxiety tax
the gap, shrinking each release
Long enough to stop being a project and become how the quarter runs. Each release lands and stays landed.
4 · Changelog
- [v0.1.0]Day 25 · 25%Added
You're tracking accomplishments. You're noticing that anxiety predictions don't come true. You're sleeping better (anxiety quieter at night). Catastrophizing is down 50%.
- [v0.5.0]Day 60 · 60%Changed
A rough week ships through the protocol intact — the bug, the terse 1:1, the missed estimate. The alarm fires, you diff it against the evidence, it resolves to noise.
- [v1.0.0]Day 100 · 100%Stabilized · stable release
You trust your skill more. Anxiety spikes but you counter with evidence. You've identified your anxiety triggers (code review, manager meeting, shipping). You have protocols for each.
Three implementers, three incident histories, one spec. Names rotated; roles and deltas real.
5 · Conformance
I shipped clean for two years and still read every review as the one that exposes me. The evidence store was the first thing that argued back with data instead of vibes.
The Friday retro is the only ritual that survived a re-org and an on-call rotation. I diff what anxiety predicted against what happened. It is wrong almost every single time.
A prod incident no longer reads as “career over.” It reads as a ticket. Same brain, recalibrated thresholds. That is the whole product, and it shipped.
Conformance note: this is a self-directed protocol, not a substitute for diagnosis, therapy, or medication. If anxiety is severe — panic attacks, avoidance — add a clinician; the spec is designed to run alongside one.
All pre-cached. We will not pretend we have not fielded each of these a few hundred times in review.
A · Frequently Raised Objections
A.1Isn't impostor syndrome just part of being an engineer?+
It is common. But common doesn't mean healthy. High-performers often have it worse. The problem is when it blocks you from taking risks or accepting promotions. That's when we treat it.
A.2If I stop feeling anxious, will I stop trying hard?+
Opposite. You'll try smarter, not harder. Right now anxiety is making you work in panic mode. Calm allows actual strategic thinking and better code.
A.3What if my anxiety is telling me something true? What if I'm not actually good enough?+
Then your track record would show that. Pull up your code reviews, your projects, your impact. If they're strong, anxiety is lying. If they're weak, we address the actual skills gap, not the anxiety.
A.4Can I do this without therapy?+
This protocol works well for mild-moderate anxiety. If anxiety is severe (panic attacks, avoidance), therapy is worth adding. But many engineers respond really well to the evidence-based protocol alone.
Merge the patch.
A 100-day protocol that turns “this is the bug that ends my career” back into a ticket that gets closed. The quiz takes two minutes, maps the events your threat model over-fires on, and hands back a 100-day spec — no shame copy, no upsell mid-quiz, no credit card.
status: PROPOSED STANDARD · ready to adopt
- A baseline read of your threat model, no devices required
- The daily reset — under two minutes per spike
- The evidence store + the weekly anxiety-vs-reality diff
- Money-back if your false-alarm rate hasn’t moved